framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,2560,0.006265777680608962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,5120,0.03594577643606398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,16384,0.057182219293382436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,2048,0.03977600071165297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,3072,0.014574221438831754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,10240,0.013439111411571503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,5120,0.017525333497259352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,2048,0.007625777688291337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,512,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,10240,0.0514373348818885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,256,0.0027644443843099806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,512,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,3072,0.2262399991353353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,128,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,512,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,768,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,1024,0.004810666872395409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,3584,0.007551111280918121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,6144,0.013447999954223633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,2048,0.006704000135262807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,2560,0.011751111182901593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,1536,0.00629066675901413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,2048,0.007424888511498769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,10240,0.03309422069125705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,3072,0.010982222027248807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,6144,0.02294844388961792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,2560,0.007312888900438945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,8192,0.008535111116038429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,256,0.005583999885453119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,3584,0.005468444277842839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,1024,0.011390222443474663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,65536,2.3569172753228083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,5120,0.010880000061459012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,10240,0.04282933473587036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,16384,0.4286204444037543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,3584,0.01016355554262797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,7168,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,768,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,10240,0.05740088886684842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,2560,0.015201777219772339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,128,0.006599999964237213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,128,0.014848889576064216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,16384,0.015328889091809591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,8192,0.6503840022616917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,1024,0.01313688854376475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,2048,0.007615110940403408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,64,0.007993777592976889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,1024,0.009138666921191746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,1024,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,512,0.005063999858167437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,12288,0.012695999609099494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,256,0.005205333232879639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,10240,0.018584888842370775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,128,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,1024,0.040876444843080305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,6144,0.06925866338941786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,512,0.018627555833922494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,64,0.005541333307822545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,2048,0.006640000061856375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,4096,0.008396444221337637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,2048,0.08334310849507649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,16384,0.024752888414594863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,7168,0.010063110954231685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,256,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,4096,0.017128000656763714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,4096,0.008347555167145198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,768,0.0029164445069101122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,2048,0.00868977771864997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,10240,0.017027555240525138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,12288,0.021959111094474792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,2560,0.03050311075316535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,2048,0.020807999703619216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,128,0.005872888697518243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,64,0.007292444507280986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,3072,0.0099973330895106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,1536,0.006597333484225803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,3584,0.015820443630218506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,1536,0.16171199745602077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,1536,0.02849777870708042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,16384,0.026604443788528442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,7168,0.037669334146711565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,2560,0.019241778386963736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,1536,0.01629066632853614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,1024,0.014761777387724983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,512,0.007667555577225155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,128,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,16384,0.023578666978412207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,512,0.005535999933878581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,256,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,64,0.007652444144090016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,3072,0.012412444584899478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,65536,0.057874666319953076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,256,0.0025271111064487034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,256,0.020939555433061387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,64,0.009034666750166152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,65536,0.08008088668187459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,8192,0.012807110945383707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,32,0.023981332778930664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,4096,0.004371555729044808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,3072,0.06144711044099596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,2560,0.0322693321439955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,512,0.004965333475006951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,512,0.003571555432346132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,2560,0.009184888667530483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,5120,0.0155102229780621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,2560,0.08253333303663465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,5120,0.02070488863521152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,1536,0.00424177779091729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,12288,0.029986666308508977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,3584,0.017902221944597032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,128,0.008987555901209513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,5120,0.06221955352359348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,5120,0.05214755402670967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,2048,0.012368888490729861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,64,0.0022026666750510535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,16384,0.08662044339709812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,3584,0.008277333445019191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,5120,0.004905777672926585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,1024,0.013461333182122974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,256,0.013758222262064615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,3072,0.02574311031235589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,10240,0.01888444523016612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,5120,0.02090399960676829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,512,0.00795911086930169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,1024,0.0028391112056043413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,2560,0.015796444482273526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,8192,0.07044621970918444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,1024,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,1536,0.026509332987997267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,16384,0.015127999915017022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,128,0.009725333087974125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,1536,0.015405333704418607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,8192,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,256,0.0022124445272816554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,5120,0.010408889088365767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,3072,0.008999110923873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,65536,0.08922666973537868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,65536,0.06018400192260742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,32,0.0028106667515304354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,7168,0.01273333364062839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,64,0.006256000035338932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,1024,0.031854222218195595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,10240,0.053596443600124784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,128,0.014728888869285583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,256,0.0024782222592168385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,2048,0.007257777783605788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,12288,0.049647112687428795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,3072,0.032207111517588295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,512,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,5120,0.015152000718646579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,768,0.014351111319330005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,1024,0.021535111798180476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,1536,0.018068444397714403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,16384,0.015506666567590503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,6144,0.14309333430396184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,2560,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,768,0.005882666756709416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,3072,0.05319199959437052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,8192,0.23972622553507486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,5120,0.015830222103330824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,7168,0.02983733349376255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,32,0.0021742222209771476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,16384,0.015301333533393012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,1536,0.017204445269372728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,5120,0.009310222334331935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,3584,0.009673777553770278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,4096,0.03198400139808655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,32,0.0018097776919603348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,10240,0.06899999909930758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,64,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,3584,0.07415733072492811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,32,0.013616888887352414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,6144,0.01275466630856196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,128,0.0058702222175068325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,3072,0.01511111193233066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,2560,0.018598222070270114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,12288,0.023026666707462732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,1024,0.01422933406300015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,2560,0.025391111771265667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,10240,0.034099555677837796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,8192,0.02108000053299798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,2048,0.011138666835096149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,65536,0.15893155998653835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,32,0.006582222051090664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,10240,0.03453777896033393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,5120,0.039878222677442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,16384,0.02509422269132402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,3584,0.022985777921146814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,7168,0.025063110722435847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,128,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,256,0.011044444309340583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,2560,0.014436443646748861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,2048,0.017143999536832173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,65536,0.07556000020768908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,2560,0.025555556019147236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,128,0.004926222066084544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,2048,0.006603555546866522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,4096,0.007636444436179266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,3584,0.009504888620641496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,2048,0.02496355606449975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,12288,0.0659102201461792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,5120,0.014464888307783337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,65536,0.05954044395022922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,2560,0.012050666742854647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,6144,0.05268266797065735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,4096,0.011049778097205691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,8192,0.041618665059407554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,1536,0.005950222412745158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,12288,0.013800000150998434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,512,0.0163128889269299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,32,0.002163555473089218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,512,0.002855111120475663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,10240,0.1071395542886522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,7168,0.009709333380063375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,4096,0.017743999759356182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,7168,0.016063110695944894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,128,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,512,0.005283555636803309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,65536,0.06308711237377591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,256,0.0114720000161065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,256,0.008995555341243744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,65536,0.05619555711746216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,7168,0.010775999890433418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,16384,0.006643555644485686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,512,0.01237511138121287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,64,0.006558222075303395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,32,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,1024,0.01751733322938283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,512,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,3072,0.014532445205582513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,3072,0.018579555882347953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,10240,0.011369778050316704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,768,0.008353778057628209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,3584,0.02702222267786662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,10240,0.022071111533376906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,7168,0.01738222274515364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,2048,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,7168,0.014097778333557976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,65536,0.07457511292563544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,4096,0.008302222523424361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,256,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,256,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,10240,0.017549332645204332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,2048,0.005948444621430502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,768,0.01684799955950843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,5120,0.01516177753607432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,512,0.10575644175211589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,1536,0.0042053332759274375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,10240,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,3584,0.016022221909628976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,2048,0.006577777779764599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,2048,0.006232000059551663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,64,0.014498665928840637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,12288,0.03611022233963013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,512,0.014873777826627096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,256,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,16384,0.07406666543748644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,8192,0.013216888739003075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,16384,0.08394310871760051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,256,0.03397688931889004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,128,0.0022071111533376905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,10240,0.016524444023768108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,16384,0.048080888059404164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,768,0.0025066667132907445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,2048,0.022983110613293115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,64,0.0028657778683635923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,768,0.006639110959238476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,6144,0.016908443636364408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,16384,0.07632178068161011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,2560,0.009325332939624786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,2560,0.0042257776690853965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,4096,0.018566222654448617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,7168,0.010032888915803697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,6144,0.014209777116775513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,6144,0.025157334076033697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,2560,0.01036800030204985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,8192,0.019912888606389362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,3072,0.022452443838119507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,2048,0.027764444549878437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,4096,0.007703111403518253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,16384,0.025069332785076563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,512,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,768,0.007756444315115611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,4096,0.0075253330998950535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,256,0.01791822248035007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,2048,0.011253332926167382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,1024,0.0031866667171319327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,1536,0.0510115557246738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,10240,0.1351333326763577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,2560,0.014133334159851074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,3072,0.017010667257838778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,512,0.01240355521440506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,2048,0.005687110953860813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,2560,0.0380364457766215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,512,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,1024,0.01240711079703437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,65536,0.1032079988055759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,8192,0.1746879948510064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,2560,0.013574221895800697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,3584,0.01331822243001726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,5120,0.00848444468445248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,2048,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,768,0.005383111122581694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,512,0.3578195571899414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,8192,0.013973333769374423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,3072,0.11430577437082927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,3072,0.09108799695968628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,512,0.005544888890451855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,3072,0.0042364444169733255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,512,0.016921778519948322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,6144,0.03538666831122504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,32,0.006611555400821898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,16384,0.018907555275493197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,12288,0.026712889472643535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,4096,0.010595555934641095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,65536,0.12691199779510498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,512,0.002942222274012036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,12288,0.028267555766635474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,2560,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,16384,0.04731200138727824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,6144,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,3072,0.004237333519591225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,6144,0.044857776827282377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,5120,0.20582489172617593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,1024,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,64,0.00657155571712388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,16384,0.023405333360036213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,16384,0.04301866557863024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,3584,0.012467555701732635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,1536,0.01683111157682207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,1536,0.004405333350102107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,1024,0.018248889181349013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,3072,0.08332266410191853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,65536,0.1541208823521932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,65536,0.11992711491054958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,1536,0.0058711109062035876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,1024,0.005616000129116907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,2560,0.019536000159051683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,12288,0.033533334732055664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,3584,0.014478221535682678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,4096,0.007622222105662028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,8192,0.011711111499203576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,65536,0.05304444498485989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,128,0.014451555079884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,12288,0.07394133011500041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,32,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,768,0.005660444498062134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,10240,0.017210667332013447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,1536,0.015559110376569959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,8192,0.042075554529825844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,512,0.01687733332316081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,512,0.01809511085351308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,10240,0.12071911493937175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,2048,0.006346666564544042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,8192,0.02917333443959554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,256,0.0069164443347189165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,10240,0.030135999123255413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,1536,0.5293777783711752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,3584,0.07902310954199897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,65536,0.7120533519321017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,2048,0.058872891796959766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,512,0.016183111402723525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,7168,0.011725333001878528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,512,0.028804444604449805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,65536,0.11955555280049641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,6144,0.019374221563339233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,1024,0.014843554960356818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,128,0.013452444639470843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,1024,0.0035422220826148987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,2560,0.06251022550794813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,2048,0.006288888967699475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,256,0.009476444787449306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,32,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,1536,0.006252444452709622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,7168,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,5120,0.01274577776590983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,12288,0.0073724447025193115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,3072,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,2560,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,65536,0.08699111143747966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,3072,0.037271999650531344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,12288,0.06647644440333049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,3072,0.007258666886223688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,64,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,1024,0.012747555143303342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,512,0.004209777961174647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,3072,0.008814222282833522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,3072,0.009307555854320526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,65536,0.2626871003044976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,16384,0.030128889613681372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,10240,0.03579911258485582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,2048,0.020245333512624104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,8192,0.06590400139490764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,256,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,128,0.001965333397189776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,6144,0.028174221515655518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,10240,0.01422933406300015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,1024,0.016586666305859882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,128,0.03103111187616984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,4096,0.0064222220745351576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,3584,0.006917333437336816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,128,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,1024,0.0207022229830424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,1024,0.01682844426896837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,256,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,256,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,256,0.002163555473089218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,10240,0.03911555475658841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,10240,0.01994311147265964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,2048,0.024012444747818842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,512,0.06528088781568739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,3072,0.0074417781498697065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,64,0.0019333333604865605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,5120,0.02072177827358246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,16384,0.017238222890430026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,8192,0.025041777226659987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,768,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,64,0.0028977776981062363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,3072,0.012348444097571902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,768,0.01683999929163191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,8192,0.17496800422668457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,5120,0.01262399968173769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,65536,0.07661510838402642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,10240,0.012681778106424542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,512,0.0024826667375034755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,1536,0.022651554809676275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,4096,0.012818666795889536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,8192,0.06603999932607015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,768,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,6144,0.09900089104970296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,65536,0.10651733477910359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,1024,0.005619555711746216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,1024,0.007639110916190677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,1024,0.13971643977695042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,3072,0.015153777268197803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,16384,0.08618044190936619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,64,0.0019004444281260173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,5120,0.052763554784986705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,3584,0.010991111397743225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,10240,0.012081777883900536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,1536,0.010043555663691627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,768,0.002586666701568498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,16384,0.031127111779318914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,8192,0.03291199935807122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,2048,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,10240,0.04115022222201029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,1024,0.01584533353646596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,5120,0.07562488979763456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,65536,0.07648799816767375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,2560,0.014484445254007975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,1536,0.009752000371615091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,128,0.004912000149488449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,7168,0.013077333569526672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,10240,0.01239822225438224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,8192,0.010363555616802638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,3584,0.01678222252262963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,512,0.005336888962321811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,512,0.01683822274208069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,65536,0.14803200297885472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,128,0.003152888889114062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,16384,0.015507555670208402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,2560,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,8192,0.00846400029129452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,10240,0.017663111289342243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,768,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,256,0.004926222066084544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,6144,0.02137511140770382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,1024,0.005756444401211209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,5120,0.11629688739776611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,2560,0.008656889200210571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,65536,0.02075377768940396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,10240,0.035028444396124944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,4096,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,10240,0.009663110805882348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,256,0.008982222113344405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,7168,0.038110223081376814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,65536,0.0784826676050822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,1024,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,8192,0.028485334581798975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,1536,0.007300444775157505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,512,0.007648888561460707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,5120,0.012915555801656512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,768,0.0028435554769304064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,128,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,256,0.016167110866970487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,256,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,5120,0.005635555419656966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,12288,0.0370764434337616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,65536,0.07402577665117052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,1024,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,12288,0.03736622134844462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,3072,0.010374222364690568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,10240,0.07148622141944037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,8192,0.013682666752073499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,256,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,2560,0.008828444613350762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,3072,0.00998577806684706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,64,0.005972444597217772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,65536,0.05772711171044243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,65536,0.03115733464558919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,1024,0.012596444123321109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,4096,0.010845333337783813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,12288,0.004978666702906291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,256,0.004334222111437055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,256,0.005697777701748743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,3072,0.03278577658865187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,65536,0.08791910939746433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,3072,0.007623111208279927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,768,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,2048,0.028484443823496502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,65536,0.18578488296932646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,768,0.0029031110720502008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,1536,0.00453333349691497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,7168,0.03204177816708883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,10240,0.03898488812976413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,3584,0.02536800007025401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,10240,0.012042666474978128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,512,0.006239111224810283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,2560,0.0337137778600057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,6144,0.09554400046666463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,7168,0.035247110658221774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,65536,0.07541600200865003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,2048,0.00796088907453749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,128,0.002283555558986134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,65536,2.6511253780788846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,64,0.0021191111041439907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,8192,0.018788443671332467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,2048,0.023312888211674158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,32,0.003043555551105075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,768,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,5120,0.013195555243227216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,3584,0.010606221854686737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,65536,0.029550221231248643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,3584,0.0188719994492001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,3584,0.024991111622916326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,7168,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,65536,0.07928266790178087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,128,0.006311111152172089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,4096,0.008255110846625434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,2048,0.006683555742104848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,128,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,4096,0.043551110559039645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,6144,0.01724177764521705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,3072,0.009314667019579146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,2048,0.012431999875439538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,4096,0.027710222535663184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,32,0.0018035556293196147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,3072,0.030472887886895075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,12288,0.02573866645495097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,7168,0.29056088129679364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,32,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,1024,0.005609777652555042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,512,0.011060444845093621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,16384,0.025428444147109985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,512,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,16384,0.050719112157821655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,5120,0.023307555251651343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,3072,0.034817778401904635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,7168,0.029176000091764662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,7168,0.034958223501841225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,2560,0.029506666792763606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,12288,0.012674666941165924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,6144,0.016198222835858662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,512,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,7168,0.34351287947760684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,8192,0.04009777638647292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,6144,0.013556444810496436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,1024,0.0028106667515304354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,7168,0.1096915536456638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,12288,0.010712888505723743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,12288,0.06290222538842095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,5120,0.04761955473158094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,12288,0.07968978087107341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,12288,0.040531555811564125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,8192,0.006341333190600078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,3584,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,4096,0.008635555704434713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,5120,0.02089777754412757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,5120,0.006346666564544042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,7168,0.03225866622394986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,3584,0.010663111176755695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,768,0.0028773333049482773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,2560,0.02847466535038418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,4096,0.02335111134582096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,4096,0.009005332986513773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,6144,0.06787466340594821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,8192,0.13038933277130127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,3072,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,2560,0.016971555021074083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,2560,0.015933333171738517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,256,0.004929777648713854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,7168,0.01723288827472263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,4096,0.00792622235086229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,65536,0.14615733093685573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,256,0.004907555464241239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,6144,0.017512000269360013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,1024,0.03613955444759793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,65536,0.1662177774641249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,2048,0.007868444754017724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,256,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,10240,0.011409777734014722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,64,0.006000888844331105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,5120,0.023323555787404377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,65536,0.055974225203196205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,512,0.015017777681350708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,64,0.0031742221779293488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,65536,0.0754222207599216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,1536,0.012072888513406118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,7168,0.030897776285807293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,2048,0.009927111367384592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,1024,0.01239111108912362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,65536,0.06733777787950304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,2048,0.0779706637064616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,4096,0.018434666925006442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,16384,1.0568826463487413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,768,0.01718844473361969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,128,0.01035644445154402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,512,0.005623999983072281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,5120,0.02919644448492262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,2560,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,12288,0.027944889333513048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,512,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,768,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,2560,0.004908444566859139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,65536,0.21792799896664092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,3072,0.06095200114780002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,512,0.0032346666687064697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,8192,0.011719110939237805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,7168,0.06649866369035509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,64,0.02786488996611701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,512,0.013763555222087435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,5120,0.012255999777052136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,12288,0.026287110315428838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,3072,0.023885332875781592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,10240,0.16080710622999403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,128,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,5120,0.008708444734414419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,5120,0.26693956057230633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,2048,0.07986666758855183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,256,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,4096,0.00794577764140235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,5120,0.009391999906963771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,7168,0.005596444424655702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,6144,0.016199999385409884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,3584,0.007618666523032718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,4096,0.012351999680201212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,65536,0.09849688741895889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,1536,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,10240,0.029888000753190782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,1024,0.004237333519591225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,7168,0.014472888575659858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,2560,0.013827555709415011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,6144,0.014360888136757744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,1536,0.027740443746248882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,768,0.003928889003064898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,64,0.006561777657932705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,16384,0.046562665038638644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,256,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,16384,0.023405333360036213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,7168,0.05159199900097317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,8192,0.24925867716471353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,3584,0.01087288889620039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,512,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,7168,0.0207324441936281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,1024,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,7168,0.05228977733188205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,5120,0.02349866595533159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,1024,0.005608888963858287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,64,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,65536,0.589280022515191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,768,0.026478222674793665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,10240,0.010904889139864178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,3072,0.009753777749008602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,3584,0.02066044509410858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,2560,0.011402666568756104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,12288,0.019073777728610568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,768,0.016539555456903245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,5120,0.008364444805516137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,128,0.0024622221373849446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,32,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,6144,0.009525333013799455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,64,0.0017928888814316855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,1024,0.0032844444115956626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,1024,0.006981333096822103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,3072,0.030322667625215318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,3072,0.0544071098168691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,10240,0.030305776331159804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,1536,0.005929777605666055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,4096,0.01738755570517646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,2048,0.012434666355450949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,768,0.005604444278611078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,768,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,6144,0.013110222087966071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,12288,0.021985777550273474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,65536,0.08131111330456205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,65536,0.7438835567898221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,1024,0.003560888684458203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,768,0.01587733295228746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,8192,0.013859555953078799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,7168,0.014177777700954013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,256,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,3584,0.018713777263959248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,8192,0.024297777149412367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,1024,0.01512977812025282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,2048,0.009770666559537252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,64,0.002177777803606457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,3072,0.005867555737495422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,8192,0.029783109823862713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,256,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,8192,0.0413973331451416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,3072,0.0069822221994400024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,6144,0.025527111358112756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,3584,0.025860443711280823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,6144,0.03122488988770379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,4096,0.014481777946154276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,3584,0.01958311100800832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,512,0.007090667055712805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,768,0.0028942223224374983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,128,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,2048,0.011128889189826118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,2560,0.020789333515697055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,3584,0.007294221884674496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,6144,0.009333333207501305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,256,0.008120888637171851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,2048,0.010324444207880232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,5120,0.01274400038851632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,3072,0.007630222373538547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,128,0.006579555571079254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,64,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,12288,0.02678488857216305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,3072,0.02608888844648997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,8192,0.16741511556837294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,256,0.013431111143694984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,6144,0.012663111090660095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,1024,0.013815110756291283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,2560,0.010410666465759277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,768,0.028845333390765723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,1536,0.008877333667543199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,128,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,6144,0.005576000031497743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,1536,0.008086221913496653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,512,0.0028986665937635633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,256,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,1024,0.014166222678290473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,512,0.006578666468461354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,2560,0.01683466633160909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,512,0.013826666606797112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,512,0.006266666783226862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,12288,0.02035999960369534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,4096,0.008658666577604082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,10240,0.01204533295498954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,5120,0.006322666588756774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,256,0.002521777732504739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,16384,0.02139288849300808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,1536,0.008752889103359645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,1024,0.003165333428316646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,5120,0.044590221510993115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,65536,4.447971343994141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,2560,0.007735999921957652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,8192,0.006278222219811545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,2048,0.06991555293401082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,64,0.00943911075592041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,32,0.02334044377009074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,10240,0.00758844448460473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,256,0.005341333233647876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,12288,0.12109511428409153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,3584,0.015790222419632804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,8192,0.139500445789761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,4096,0.021966222259733412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,7168,0.03153511219554477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,7168,0.07906399832831489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,8192,0.011029333704047732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,512,0.011741333537631564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,1536,0.00904266701804267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,512,0.01384800010257297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,65536,0.025088888075616624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,12288,0.017264889346228707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,1536,0.016200888488027785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,6144,0.009540444446934594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,32,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,6144,0.011708444191349877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,512,0.006650666809744305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,1536,0.006239999913507038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,128,0.004876444323195351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,6144,0.00794933322403166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,64,0.0018168888572189545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,768,0.01126844435930252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,4096,0.016367110941145156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,512,0.005977777971161737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,5120,0.009705777797434065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,2560,0.03239733311865065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,256,0.00796888851457172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,128,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,12288,0.028189334604475234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,2048,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,8192,0.008975110948085785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,2048,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,768,0.006548444430033366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,64,0.005622222191757626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,16384,0.023524444964196947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,7168,0.03109955456521776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,32,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,10240,0.01717422240310245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,3584,0.017200888858901132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,32,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,3584,0.013423110875818463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,3072,0.10133777724372016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,3072,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,32,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,7168,0.010036444498433007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,5120,0.423999998304579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,768,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,128,0.018256000346607633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,128,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,8192,0.28532444106207955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,1536,0.011010666688283285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,1536,0.006301333506902059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,2560,0.010666666759385003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,5120,0.012716444830099741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,4096,0.016487111647923786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,5120,0.01752000053723653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,64,0.005952888892756567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,2560,0.019876443677478366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,7168,0.019710222880045574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,1536,0.005945777727497949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,128,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,4096,0.007910221815109253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,2048,0.00962577760219574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,768,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,256,0.0107022225856781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,65536,0.07653866873847114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,2560,0.007992000215583378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,65536,0.09580710861417983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,3072,0.007296000089910295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,512,0.008339555727110969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,5120,0.008287111090289222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,12288,0.025340444511837427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,2560,0.007812444534566667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,2560,0.05204000075658163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,16384,0.7043271064758301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,7168,0.012453333371215396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,12288,0.031425777408811785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,2560,0.03181511163711548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,256,0.03195377853181627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,8192,0.30102843708462185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,128,0.00798044436507755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,1024,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,65536,0.36139289538065594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,32,0.005535999933878581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,7168,0.014262222581439547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,1024,0.0029066666546795103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,8192,0.015776000089115567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,5120,0.008767111433876885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,3584,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,2560,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,64,0.00795111142926746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,12288,0.03900711072815789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,2560,0.02677511175473531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,128,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,32,0.00250133333934678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,10240,0.021485333641370136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,3072,0.005960888746711943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,3072,0.007373332977294922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,8192,0.01956977778010898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,3584,0.015124445160230001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,10240,0.015645333462291293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,7168,0.009999111294746399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,12288,0.020616888999938965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,65536,0.11370400587717693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,5120,0.010634666515721215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,1536,0.0031724443866146933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,2560,0.018961777289708454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,3072,0.008328888979223039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,10240,0.04386933313475715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,768,0.01240888900227017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,16384,0.02516977820131514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,65536,0.0729057788848877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,256,0.016213332613309223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,7168,0.01753777762254079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,7168,0.03402400016784668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,16384,0.015523556205961438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,1536,0.014835556348164877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,3072,0.02776355544726054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,7168,0.06681778033574422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,768,0.0031439999325407874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,12288,0.2200533284081353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,4096,0.023303111394246418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,8192,0.03390844331847297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,768,0.029695110188590154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,8192,0.03276177909639146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,5120,0.021104888783560857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,768,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,512,0.01352266636159685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,128,0.005534222142563925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,2048,0.006064888917737537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,512,0.006211555666393704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,65536,0.08474666542477077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,128,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,12288,0.0195795562532213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,768,0.016568889220555622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,10240,0.01553422212600708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,8192,0.11603821648491754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,16384,0.07877422041363187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,6144,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,3072,0.022044445077578228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,7168,0.005203555441564984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,8192,5.340361701117621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,128,0.007260444263617198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,3072,0.019871110717455547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,256,0.014131555954615274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,3584,0.006931555353932911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,512,0.01277599980433782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,768,0.0026791110220882627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,1536,0.026055110825432673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,512,0.002480000050531493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,2560,0.0041991112132867174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,3584,0.011468444433477191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,1024,0.048150221506754555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,1024,0.014436443646748861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,5120,0.03947822252909342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,768,0.06436089012357923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,256,0.013387555877367655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,2048,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,3584,0.009117333425415887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,10240,0.011358222199810876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,2560,0.011360888679822287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,768,0.007626666790909237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,1536,0.016711999972661335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,65536,0.07451733615663317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,16384,0.07101955678727892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,256,0.013786666923099093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,256,0.0022213332768943575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,16384,0.02026844521363576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,1536,0.00351200004418691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,16384,0.02433066732353634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,5120,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,4096,0.008002666963471307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,16384,0.026752000053723652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,10240,0.04289422101444668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,2048,0.011914666328165265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,5120,0.008301333420806462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,3584,0.018960000740157235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,10240,0.011718221836619906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,128,0.01513777838812934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,1536,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,7168,0.02905155552758111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,10240,0.011718221836619906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,512,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,10240,0.01718577742576599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,512,0.005541333307822545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,7168,0.030825777186287776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,1536,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,768,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,256,0.06773244672351413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,65536,0.13053154945373535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,768,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,256,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,1536,0.004567999806669024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,128,0.015811555915408664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,4096,0.011012444065676795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,6144,0.009444444543785518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,65536,0.14386755890316433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,256,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,32,0.0018586667461527719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,512,0.002512000087234709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,256,0.011807999677128263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,1536,0.005954666684071223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,16384,0.05735288725958931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,1024,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,1024,0.005919111271699269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,3584,0.020065777831607394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,16384,0.06580177942911784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,2560,0.010011555420027839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,64,0.006538666784763336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,12288,0.011216888825098673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,1024,0.018702222241295707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,512,0.009331555830107795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,3072,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,768,0.01722311145729489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,2048,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,2560,0.00453333349691497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,128,0.007617777420414819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,7168,0.07194666730033027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,768,0.016857778032620747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,6144,0.017483555608325534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,12288,0.05457333392567105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,7168,0.026567111412684124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,2048,0.008998221821255153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,6144,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,256,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,2048,0.013415999710559845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,10240,0.07225333319769965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,12288,0.09052799807654487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,6144,0.01533066729704539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,16384,0.15826222631666395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,1536,0.00996088898844189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,5120,0.00462400002612008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,128,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,10240,0.20900000466240776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,768,0.017516444126764934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,256,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,5120,0.005000000198682149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,3584,0.01936177743805779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,256,0.014480888843536377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,3584,0.005946666830115848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,3584,0.0063688887490166565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,768,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,1024,0.006220444622966979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,2560,0.004911111046870549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,6144,0.0329146683216095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,7168,0.02658933401107788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,2048,0.006312888943486744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,128,0.0029146667155954572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,16384,0.21033244662814668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,3584,0.007599111232492659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,64,0.003137777869900068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,64,0.002141333288616604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,3072,0.007302222152551015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,768,0.005790222022268508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,12288,0.061448891957600914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,7168,0.045223110251956515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,128,0.012438221938080259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,3584,0.0684373312526279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,256,0.014132445057233175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,16384,0.012368000215954252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,1024,0.02780533168050978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,32,0.0018533333722088074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,512,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,2048,0.013381332986884646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,5120,0.01143022212717268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,16384,0.017031111650996737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,12288,0.01965777741538154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,2560,0.00848444468445248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,768,0.0064702220261096954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,3584,0.012329777909649743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,768,0.017312000195185345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,2560,0.016354666815863717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,512,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,16384,0.020580444071027968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,32,0.006206222292449739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,3584,0.007296889192528195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,65536,0.2428426742553711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,10240,0.21866844760047066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,256,0.00553955551650789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,1024,0.004927111168702443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,1024,0.01583377851380242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,1024,0.024119999673631456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,2048,0.021710221966107685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,256,0.005938666562239329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,6144,0.009686222506894005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,65536,0.08466044399473403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,32,0.0019102222803566193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,16384,0.13289599948459202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,65536,0.5762400097317165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,32,0.0793902211719089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,16384,0.00867555538813273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,3072,0.00656355544924736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,768,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,1024,0.033562666840023465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,128,0.01715022159947289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,12288,0.04753066764937507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,2560,0.017913776967260573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,6144,0.024004444479942322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,128,0.01332266628742218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,768,0.016213332613309223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,1024,0.0031271111220121384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,5120,0.013415999710559845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,4096,0.007602666815121968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,1024,0.00314666661951277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,6144,0.019870221614837646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,256,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,65536,0.12294222248925103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,768,0.0038968887594011095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,6144,0.011732444167137146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,1536,0.00977600034740236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,2048,0.04345422320895725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,512,0.00591111100382275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,16384,0.050256000624762646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,1024,0.0031600000543726813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,10240,0.040603554911083646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,1024,0.014866666661368476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,4096,0.007962666451931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,64,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,65536,0.07640000184377034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,32,0.006221333311663733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,3072,0.017252445220947266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,2560,0.019581332802772522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,512,0.014984000060293408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,128,0.002136888810329967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,7168,0.00963466697269016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,4096,0.00777511133088006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,3072,0.012552888856993781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,2048,0.012695111334323883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,65536,0.1197555594974094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,4096,0.010563555690977307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,1536,0.006450666735569636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,5120,0.008287111090289222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,65536,0.12165066930982803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,12288,0.04077333211898804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,64,0.005321777943107817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,5120,0.008650666309727563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,10240,0.018613333503405254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,1024,0.016922665966881644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,2560,0.006707555717892117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,3072,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,1024,0.007547555698288812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,4096,0.005937777873542573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,5120,1.5659422344631617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,3584,0.011144888897736868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,512,0.014487110906177096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,5120,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,65536,0.08481155501471625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,7168,0.024832000335057575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,1024,0.004295999805132548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,12288,0.07121866941452026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,32,0.006265777680608962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,5120,0.12303466267055935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,12288,0.02590755621592204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,16384,0.3639049000210232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,5120,0.09360000160005356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,1536,0.00831911133395301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,3072,0.01484444406297472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,4096,0.01071999967098236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,12288,0.04733955529001024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,2048,0.011752000285519494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,7168,0.008807111117574904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,12288,0.007425777614116669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,16384,0.032524443335003324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,1024,0.00619911112719112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,12288,0.020152888364262052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,6144,0.01313777764638265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,2560,0.004922666483455234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,1024,0.005944889038801193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,65536,0.05684533384111193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,64,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,6144,0.019285332825448778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,2560,0.04724355538686117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,2048,0.006611555400821898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,512,0.007962666451931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,5120,0.016559110747443307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,16384,0.016574222180578444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,3584,0.020602666669421725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,12288,0.013993778162532382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,128,0.0018088887963030075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,64,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,4096,0.04923911227120293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,6144,0.032884445455339216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,6144,0.013903111219406128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,64,0.001847111102607515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,8192,0.019879110985332064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,1024,0.006278222219811545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,12288,0.023056000471115112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,64,0.0035662220584021676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,2048,0.006321777900060018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,65536,0.12830489211612278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,128,0.004942222187916438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,8192,0.01460622251033783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,768,0.007244444555706448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,1024,0.005618666609128316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,2048,0.006578666468461354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,7168,0.012654222548007965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,1024,0.006212444355090459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,768,0.005577777822812398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,256,0.008282666405042013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,1536,0.03285600079430474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,4096,0.010238221950001186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,6144,0.010688888529936472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,4096,0.01567911108334859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,2048,0.05221422182189094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,256,0.014451555079884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,3072,0.017299556069903903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,512,0.02398666739463806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,65536,0.07762844694985284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,3584,0.007269333634111617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,2048,0.22360089090135363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,7168,0.012513777448071374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,1024,0.005662222289376789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,12288,0.3454577657911513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,16384,0.02611200014750163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,5120,0.030028445853127375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,12288,0.017064889272054035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,2560,0.16432532999250624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,16384,0.03403110967742072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,768,0.029755555921130713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,8192,0.014780445231331719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,128,0.0022337778160969415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,65536,3.217448976304796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,32,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,3584,0.06028978029886881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,10240,0.02404888967672984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,8192,0.02442844377623664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,10240,0.07216978073120117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,12288,0.004964444372389051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,32,0.0022320000247822865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,8192,0.012137778103351593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,4096,0.008035555481910706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,2560,0.007335999773608313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,8192,0.04110577702522278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,3584,0.010999110837777456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,256,0.002858666703104973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,3072,0.015830222103330824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,512,0.0038604442444112566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,3584,0.009848888549539778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,64,0.002151111140847206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,6144,0.008989333278603023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,7168,0.00942755573325687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,10240,0.011943110989199745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,2560,0.022343110707071092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,128,0.0063457778758472866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,2048,0.006694222076071634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,16384,0.024041778511471216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,1536,0.011244444383515252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,2048,0.006405333264006509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,128,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,1536,0.021723555194007024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,5120,0.11311554908752441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,1024,0.2231395509507921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,128,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,128,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,2048,0.03626399901178148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,12288,0.008794666992293464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,7168,0.015456888410780164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,7168,0.016899555921554565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,768,0.017937777770890128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,256,0.016196444630622864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,128,0.027979556057188246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,16384,0.04931466778119405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,5120,0.008634666601816813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,512,0.02369511127471924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,1536,0.025070221887694463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,2048,0.03665155503484938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,32,0.0019040000107553268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,7168,0.01125955581665039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,10240,0.037068443165885076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,3072,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,1536,0.006219555520349079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,5120,0.12391466564602321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,2048,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,512,0.0053173332578606075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,2048,0.050240000089009605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,12288,0.014095111025704278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,1536,0.029116445117526587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,5120,0.02748799986309475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,65536,0.08590222067303127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,2048,0.004060444318585926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,2560,0.006775110960006714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,7168,0.01568177839120229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,3584,0.014349333114094205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,128,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,4096,0.008330666356616551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,1024,0.011737777955002256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,128,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,1536,0.02267733381854163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,3584,0.027631999717818365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,3072,0.018558222386572096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,5120,0.18760977851019967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,16384,1.5809715059068468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,3584,0.022524444593323603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,10240,0.01404888927936554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,7168,0.01311644415060679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,2048,0.0079644446571668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,65536,0.07761777771843804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,256,0.00793866647614373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,65536,0.09652355644438003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,32,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,1536,0.008370666868156856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,1024,0.0041777777175108595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,1024,0.02279022170437707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,1024,0.02477688921822442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,1536,0.006102222121424145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,5120,0.027756444282001917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,2560,0.006615110983451207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,7168,0.06626044379340278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,65536,0.34308444129096133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,65536,0.12499199973212348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,32,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,6144,0.009311111436949836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,6144,0.04155288802252875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,4096,0.02578488820128971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,128,0.006586666736337874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,7168,0.04938311047024197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,65536,0.16573777463701037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,768,0.0032222221295038858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,5120,0.008653333617581261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,4096,0.06544799937142266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,768,0.0032097777972618737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,1536,0.03751022285885281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,256,0.004877333425813251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,32,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,12288,0.019698666201697457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,12288,0.015354666444990369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,6144,0.03534488876660665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,6144,0.03757777810096741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,8192,0.02838311261600918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,1024,0.014508444401952954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,16384,0.058470222685072154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,6144,0.016145777371194627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,4096,0.007292444507280986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,3072,0.025839999318122864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,16384,0.029814223448435467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,768,0.00658044425977601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,1536,0.025705777936511572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,768,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,8192,0.017695999807781644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,12288,0.018349332941903006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,768,0.0070186663005087115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,128,0.00416088890698221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,8192,0.020272889071040683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,2048,0.06022044685151842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,8192,0.012420444852775998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,65536,0.043806221750047475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,7168,0.030023998684353296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,6144,0.017180444465743173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,2048,0.011767110890812345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,2048,0.017635555730925668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,3584,0.01309511148267322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,1024,0.017703110973040264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,32,0.008660444782839881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,128,0.002152888932161861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,1024,0.007278222176763747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,64,0.0018462222069501877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,5120,0.008655110994974772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,128,0.015113777584499784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,256,0.014132445057233175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,6144,0.06639199786716037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,65536,0.10234222147199844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,64,0.0019324444648292328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,2560,0.019567999574873183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,16384,0.01790844400723775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,2560,0.013431111143694984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,768,0.0199617776605818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,1536,0.004204444587230682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,65536,0.03502488798565335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,512,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,8192,0.013074666261672974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,3584,0.08842133151160346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,7168,0.013182222015327878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,128,0.0021226666867733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,8192,0.0148044443792767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,65536,0.14856710698869494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,256,0.0022008888837363985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,6144,0.014123555686738757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,7168,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,128,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,64,0.006558222075303395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,3072,0.009371555513805812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,5120,0.012448000411192576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,768,0.00282222218811512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,2560,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,32,0.0018524444765514797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,256,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,10240,0.017312000195185345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,6144,0.020453333854675293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,10240,0.02754577828778161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,32,0.0018453333112928604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,5120,0.008979555633332994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,16384,0.06020355224609375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,10240,0.015119110544522604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,256,0.018090666996108163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,12288,0.2387324439154731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,128,0.014427555931939019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,128,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,3072,0.10185866885715061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,4096,0.012373333175977072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,128,0.008600888980759515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,1536,0.006217777729034424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,4096,0.006235555642180973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,32,0.013427555561065674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,12288,0.21810132927364775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,4096,0.007672888537247975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,128,0.005527110977305307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,512,0.003555555724435382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,128,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,1024,0.03194222185346816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,2048,0.014325333966149224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,4096,0.011858666936556498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,10240,0.1263857815000746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,8192,0.11184444692399766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,8192,0.016979555288950603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,512,0.0031840000301599503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,3584,0.007535999847782983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,6144,0.025719111164410908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,1024,0.008281778130266402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,256,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,7168,0.009122666385438707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,12288,0.01648977730009291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,65536,0.08409777614805435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,1024,0.005640888793600931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,128,0.012799999780125089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,2560,0.013420444395807056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,2560,0.1466266711552938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,8192,0.011051555474599203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,12288,0.014142221874660917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,65536,0.04972177743911743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,4096,0.02271822260485755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,2560,0.026354667213228013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,4096,0.01682222220632765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,3072,0.006059555543793573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,7168,0.06677155362235175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,4096,0.03894933395915561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,2048,0.006594666590293248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,4096,0.007975999679830339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,65536,0.084842668639289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,7168,0.013049778011110095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,512,0.005623111294375525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,16384,0.03331022130118476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,65536,0.09495555692248875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,1536,0.00980977796845966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,3072,0.008280889027648503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,3584,0.007296000089910295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,3072,0.01275199982855055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,3072,0.00699733363257514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,8192,0.09617777665456136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,1024,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,16384,0.03615555498335097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,32,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,10240,0.030119998587502375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,2560,0.005609777652555042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,3584,0.0074480002125104266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,4096,0.009791110952695211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,8192,0.015623999966515435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,1536,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,12288,0.08087555567423503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,3072,0.027446221974160936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,2560,0.009983110758993361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,512,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,6144,0.017439999514155917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,1536,0.09646488560570611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,3584,0.029559999704360962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,128,0.016840888394249808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,10240,0.024056888288921777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,6144,0.029860443539089624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,2048,0.006648889018429651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,10240,0.19675111770629883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,128,0.007658667034573025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,768,0.003559999995761447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,64,0.0018382221460342407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,16384,0.04357777701483833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,2048,0.013346666263209449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,8192,0.019870221614837646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,128,0.016872889465755887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,65536,0.1528808938132392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,65536,0.08160622252358331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,4096,0.10065422455469768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,65536,0.06451555755403307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,512,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,65536,0.09548266728719075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,6144,0.020236444142129686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,65536,3.717962476942274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,65536,0.09146933423148261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,7168,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,16384,0.031930665175120033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,3584,0.012368888490729861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,2048,0.07965244187249078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,1536,0.006138666636413998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,512,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,65536,0.0860968894428677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,5120,0.009176000124878353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,512,0.00831288927131229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,6144,0.015253333581818474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,16384,0.33185601234436035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,12288,0.030460443761613634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,3584,0.007584888901975419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,6144,0.02473955518669552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,5120,0.01092088884777493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,1536,0.018954666124449838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,1024,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,768,0.01517511076397366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,3072,0.01440711153878106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,768,0.02124711043304867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,16384,0.02814399864938524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,6144,0.29206755426194936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,65536,0.5263626840379503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,16384,0.02476177778508928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,512,0.019754666421148512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,768,0.00453599997692638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,12288,0.01983022193113963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,128,0.03129422333505418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,8192,0.012079111403889127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,65536,0.9797875086466471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,32,0.010051555931568146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,4096,0.027792887555228338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,12288,0.04498755600717333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,12288,0.004664888812435998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,3072,0.008020444048775567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,8192,0.011743110915025076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,5120,0.026520889666345384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,4096,0.010772444307804108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,2560,0.006686222222116258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,7168,0.02757599949836731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,2560,0.13575288984510633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,4096,0.012060444388124677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,2048,0.01536266671286689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,7168,0.04066933194796244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,128,0.007273777491516537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,3072,0.0068711112770769335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,2560,0.007287110719415877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,16384,0.018280888597170513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,256,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,3072,0.016921778519948322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,10240,0.025891555680169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,1536,0.017557332913080852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,65536,0.0561351113849216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,7168,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,512,0.0021297778520319196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,256,0.021729777256647747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,4096,0.030770666069454614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,6144,0.015853333804342482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,1536,0.012714666624863943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,64,0.0018577778504954444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,128,0.0028008888992998335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,128,0.002191999927163124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,3072,0.09404622183905707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,12288,0.02863466739654541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,128,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,5120,0.008880889250172509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,10240,0.015864888827006023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,3072,0.004553777890072929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,512,0.002481777841846148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,2560,0.15628711382548013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,16384,0.02455111179086897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,6144,0.01553422212600708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,2560,0.0241333329015308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,2048,0.007276444799370236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,12288,0.023726221587922838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,1024,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,2560,0.01831022236082289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,6144,0.009343999955389235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,128,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,128,0.013416888813177744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,3584,0.08444178104400635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,32,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,8192,0.03604888916015625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,2048,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,7168,0.01994400057527754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,1536,0.022986667023764715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,1024,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,5120,0.051015999582078725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,256,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,256,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,7168,0.008639111287064022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,2048,0.011630222201347351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,2560,0.04152266515625848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,3072,0.023756444454193115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,32,0.004868444469239977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,7168,0.022851554883850947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,768,0.01614488826857673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,12288,0.01719555589887831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,8192,0.028806222809685603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,4096,0.01239999963177575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,6144,0.01825066738658481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,768,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,5120,0.023560888237423364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,1024,0.014399111270904541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,4096,0.022685334086418152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,7168,0.020257777637905545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,16384,0.19446133242713082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,1024,0.006631999793979857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,1536,0.036024000909593366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,12288,0.02775555517938402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,5120,0.03299644589424133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,256,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,256,0.002377777877781126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,8192,0.07650844256083171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,2048,0.02489600082238515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,4096,0.007662222617202335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,7168,0.018497778309716117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,512,0.00250133333934678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,7168,0.023009777069091797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,256,0.020759110649426777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,8192,0.13076266977522108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,2048,0.013236444857385425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,12288,0.024167110522588093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,256,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,8192,0.014426666829321118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,16384,0.07128355238172743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,64,0.0026666666898462507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,256,0.0025048889219760895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,256,0.01408088869518704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,12288,0.05897600120968289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,16384,0.029804444975323145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,7168,0.04105510976579454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,12288,0.03345866666899787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,12288,0.016912000046836007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,16384,0.07708799839019775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,3072,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,32,0.001814222170246972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,256,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,3584,0.005630222045713001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,5120,0.2270409001244439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,8192,0.01035377797153261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,3584,0.007247111035717859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,6144,0.11607821782430013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,3072,0.012049777640236748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,256,0.002624888800912433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,64,0.0024746666765875286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,256,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,128,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,2560,0.022657778528001573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,128,0.014097778333557976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,3584,0.024688889582951862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,4096,0.00961866643693712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,256,0.006232000059551663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,12288,0.0740106635623508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,65536,0.17131111356947157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,4096,0.009872888525327047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,2560,0.006646222124497096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,4096,0.011757333245542316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,4096,0.016543999314308167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,3584,0.24908799595303008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,10240,0.011719110939237805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,512,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,2560,0.02033422225051456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,7168,0.031222220924165513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,512,0.027133334014150832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,3584,0.004880889008442561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,4096,0.04744977752367655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,3072,0.00720088928937912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,256,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,65536,0.11411200629340278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,16384,0.01995644470055898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,512,0.013184888495339287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,1536,0.007286222444640265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,65536,0.11994666523403591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,64,0.001820444439848264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,2048,0.007598222129874759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,1536,0.008283555507659912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,8192,0.06045955419540405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,768,0.011167111496130625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,65536,0.009329777624871995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,65536,0.3013022210862902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,3072,0.19244444370269775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,2560,0.007625777688291337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,1024,0.010647999743620554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,8192,0.07419555717044406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,4096,0.034625778595606484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,32,0.037017779217825994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,8192,0.014986667368147107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,3072,0.005542222410440445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,4096,0.00794666674402025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,5120,0.025634666283925373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,7168,0.09240800142288208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,3072,0.10220444202423096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,12288,0.04311733444531759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,65536,0.14200977484385172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,7168,0.03743022349145677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,10240,0.02678488857216305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,7168,0.038310223155551486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,128,0.01735733283890618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,8192,0.010416000253624385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,3584,0.01978933314482371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,7168,0.022395556171735127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,7168,0.10565066337585449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,1536,0.011062222222487131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,3584,0.04464888903829786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,3072,0.02199733257293701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,8192,0.04686133397950066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,1536,0.00630666646692488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,7168,0.008958222137557136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,5120,0.018176888426144917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,12288,0.014020444618331062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,1536,0.01737333337465922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,4096,0.005961777849329843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,12288,0.04001155495643616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,512,0.0029564443975687027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,6144,0.06918488608466254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,6144,0.024116444918844435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,7168,0.028239998552534316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,5120,0.033619556162092425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,65536,0.16359733210669622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,16384,0.023892444041040208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,1024,0.008628444539176093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,65536,0.07485600312550862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,256,0.0021395554973019496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,4096,0.12797421879238552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,3584,0.011014222270912595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,512,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,6144,0.008957333034939235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,1536,0.0314995547135671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,16384,0.015272888872358533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,7168,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,2560,0.014448000325096978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,12288,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,8192,0.04447911183039347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,256,0.015461333923869662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,65536,0.17575021584828696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,2560,0.0571848882569207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,256,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,7168,0.08671555254194473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,4096,0.06129778093761868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,5120,0.015105777316623263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,5120,0.010057777994208867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,2560,0.008279110822412703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,512,0.0210533340771993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,128,0.014817777607176038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,3584,0.010766222245163388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,1024,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,512,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,16384,0.029687109920713637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,3072,0.014045332868893942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,6144,0.043291555510626904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,256,0.004540444248252445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,2048,0.027927110592524212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,12288,0.012741333080662621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,8192,0.03459999958674113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,4096,0.007983999947706858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,3072,0.02591466738118066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,2048,0.008004444340864817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,5120,0.012438221938080259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,32,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,12288,0.05954933166503906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,1024,0.005922666854328579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,6144,0.146323561668396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,512,0.002558222247494592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,3072,0.022542221678627863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,3072,0.010803555448849997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,2560,0.00793599999613232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,16384,0.41602399614122176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,256,0.014106666048367819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,512,0.0025173332542181015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,4096,0.01551466683546702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,7168,0.009711110757456886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,2048,0.005975111077229182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,8192,0.02587022218439314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,12288,0.10869510968526204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,1536,0.022472888231277466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,3584,0.01032977799574534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,7168,0.009777777724795872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,5120,0.013328889177905189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,6144,0.02661955522166358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,12288,0.2630026605394152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,65536,0.26522665553622776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,256,0.004947555561860402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,3072,0.007888000044557784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,1024,0.012462221913867526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,1536,0.00798044436507755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,6144,0.009711110757456886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,1536,0.005924444645643234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,1024,0.005247999810510212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,2048,0.012755555411179861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,3072,0.007785778078767989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,3072,0.006572444405820634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,256,0.005205333232879639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,256,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,32,0.001793777777089013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,256,0.0033848887930313745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,5120,0.27294932471381295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,2560,0.006649777707126405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,7168,0.013979555832015144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,512,0.0069288888739215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,12288,0.034430222378836736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,6144,0.009992889232105678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,5120,0.020296888218985665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,256,0.00490133340160052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,65536,0.06058222055435181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,7168,0.04465422365400526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,3584,0.026016000244352553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,3072,0.015881778465376962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,1536,0.006585777633719974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,3584,0.010069333016872406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,5120,0.09247733487023248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,10240,0.015321777926550971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,3072,0.015232000086042615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,7168,0.010712888505723743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,12288,0.023734221855799358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,6144,0.06340089109208849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,16384,0.03747111227777269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,65536,0.06011288695865207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,16384,0.02628266645802392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,3072,0.036455111371146307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,3584,0.007481777833567725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,3584,2.3508987426757812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,4096,0.030198223061031763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,10240,0.03374400072627597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,8192,0.02071199980047014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,16384,0.07342222001817492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,2048,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,256,0.02069511181778378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,10240,0.031215111414591473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,512,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,1024,0.06572533316082425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,256,0.013058666553762225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,256,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,65536,0.16473510530259874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,1536,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,2048,0.008700444466537898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,1536,0.01759466694460975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,16384,0.07262488868501452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,3072,0.03085333440038893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,7168,0.010692444112565784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,2560,0.018241778016090393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,64,0.006072888771692912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,16384,0.07902310954199897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,65536,0.12005066871643066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,1024,0.010764444039927589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,512,0.003903999924659729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,12288,0.027800000376171533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,768,0.0028248888750871024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,1536,0.029811554484897192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,8192,0.011104889214038849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,3072,0.013996443814701505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,64,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,5120,0.01479200025399526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,65536,0.08379822307162815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,2048,0.03530044356981913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,12288,0.021703110800849065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,4096,0.044490665197372437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,6144,0.01848355597919888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,32,0.006528000036875407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,5120,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,3584,0.013041777743233575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,512,0.008277333445019191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,512,0.006909333169460297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,10240,0.020651555723614164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,2560,0.019846222466892667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,8192,0.011699555648697747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,128,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,8192,0.01477599971824222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,768,0.007263110743628607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,3584,0.013601777454217276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,3584,0.0148026661740409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,32,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,64,0.005974222388532426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,3584,0.027475555737813313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,768,0.011126221881972419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,768,0.012688000169065265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,3584,0.01148799972401725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,1024,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,7168,0.041166222757763334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,8192,0.010925333533022137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,256,0.02366666661368476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,2560,0.00693866651919153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,1536,0.006486222147941589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,768,0.01483022173245748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,2048,0.0151582227812873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,8192,0.01757244434621599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,3072,0.004552888787455029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,128,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,256,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,3072,0.009352888498041365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,4096,0.10223733054267035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,16384,0.015164444843928019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,6144,0.013457777599493662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,8192,0.020170667105250888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,256,0.00250133333934678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,16384,0.048763556612862476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,5120,0.009332444104883406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,8192,0.13890222708384195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,256,0.024329778220918443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,7168,0.021764443980322942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,5120,0.004591111093759537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,256,0.04327466752794054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,512,0.162517335679796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,7168,0.038559112283918597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,65536,0.03623644510904948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,64,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,2048,0.006656888872385025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,768,0.02165777815712823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,512,0.013752889302041797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,3584,0.010046222143703038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,12288,0.054401778512530856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,256,0.012416889270146688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,768,0.0058782220714622075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,256,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,32,0.006212444355090459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,65536,0.08000710937711927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,16384,0.05686755312813652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,1024,0.009330666727489894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,5120,0.014230221509933472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,16384,0.03052266769938999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,4096,0.017064000169436138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,12288,0.022338666849666174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,10240,0.015860444969601102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,16384,0.028055111567179363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,512,0.005505777895450592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,8192,2.6347058614095054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,3072,0.016514667206340365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,1024,0.006798222247097228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,1536,0.018057778477668762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,2048,0.012169777519173093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,768,0.03326844506793552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,1024,0.038551112016042076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,8192,0.03568266828854879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,128,0.005881777654091517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,1024,0.006324444380071428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,2560,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,1024,0.00554577757914861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,1536,0.01997155613369412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,6144,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,10240,0.03321511215633816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,4096,0.05269155568546719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,7168,0.010680889089902243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,3072,0.007852444218264686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,768,0.009975999593734741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,2048,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,16384,0.018946665856573317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,16384,0.12073866526285808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,256,0.025451555848121643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,1024,0.005886222339338726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,256,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,2048,0.006976889239417181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,16384,0.13627110587226018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,3072,0.0103511114915212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,256,0.0025315555847353404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,1024,0.014454222387737699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,2560,0.007254222200976477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,12288,0.013451555536852943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,768,0.018225777480337355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,10240,0.019796444310082328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,10240,0.2429262267218696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,2560,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,1024,0.023360000716315374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,16384,0.025880000657505457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,128,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,2560,0.020614221692085266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,1536,0.019296889503796894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,512,0.018583110637134977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,12288,0.013066666821638743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,6144,0.01976799964904785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,6144,0.21821510791778564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,7168,0.010676444404655032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,16384,0.05789688560697767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,1024,0.003173333282272021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,3072,0.02234755622016059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,1024,0.0032071111102898917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,512,0.0028088889602157804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,256,0.01239733315176434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,6144,0.004242666479614046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,1536,0.0329039990901947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,1024,0.03046311272515191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,16384,0.013759111364682516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,3072,0.009787555370065901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,1024,0.006225777582989798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,1536,0.003540444291300244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,3584,0.01329955541425281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,3584,0.007263999846246507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,4096,0.04010666741265191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,2560,0.09459733300738865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,768,0.01851022243499756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,2048,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,16384,0.09045333332485622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,768,0.027225777506828308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,64,0.002136888810329967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,2048,0.008019555773999956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,7168,0.035066667530271746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,10240,0.015476443701320224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,8192,0.011081777513027191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,10240,0.011021333436171213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,512,0.0026933333526055017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,128,0.004987555659479565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,7168,0.07030133406321208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,1536,0.03843999902407328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,768,0.0028284444577164124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,256,0.01036000003417333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,2560,0.007275555696752336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,2560,0.007314666277832455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,1024,0.030856887499491375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,128,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,8192,0.01724088854259915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,2560,0.014774221513006421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,16384,0.04167644513977898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,7168,0.014147554834683737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,10240,0.14092444048987493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,1024,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,1536,0.008833777573373582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,65536,0.16233155462476942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,512,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,3072,0.00962400022480223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,4096,0.008657777474986183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,5120,0.005608888963858287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,12288,0.013055110971132914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,256,0.004937777916590373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,4096,0.01847288840346866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,768,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,256,0.003871110992299186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,12288,0.04068888889418708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,32,0.006903111106819577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,2560,0.006688000013430913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,12288,0.03353688783115811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,2560,0.006007110906971826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,1024,0.00453333349691497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,32,0.002161777681774563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,512,0.0032275555034478507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,1536,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,5120,0.010130667024188572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,7168,0.009961778091059791
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,12288,0.10374844736523098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,32,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,4096,0.13708621925777859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,32,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,1536,0.003625777860482534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,1536,0.015170666906568738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,12288,0.16958844661712646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,512,0.002524444419476721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,2560,0.016551110479566786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,1024,0.003168888803985384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,10240,0.01183377785815133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,1536,0.0038737778862317405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,1024,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,3584,0.009692444569534725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,1536,0.003891555385457145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,128,0.002471999989615546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,5120,0.008952889177534316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,12288,0.01534222231970893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,4096,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,3072,0.017816889617178176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,2048,0.004523555437723796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,10240,0.03423644436730279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,12288,0.09762310981750488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,1536,0.003555555724435382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,5120,0.050782223542531334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,3584,0.007480888730949826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,3584,0.12959199481540257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,3584,0.007470221983061896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,2560,0.006895111252864202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,2560,0.018280888597170513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,2048,0.004228444563017951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,2560,0.010049777726332346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,128,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,32,0.020971554848882887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,2560,0.009753777749008602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,512,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,3584,0.16237510575188532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,5120,0.05096977618005541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,3584,0.007243555453088548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,5120,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,4096,0.012077333198653327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,1024,0.007650666766696506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,8192,0.018199111024538677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,2048,0.007323555648326874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,65536,0.0912968913714091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,7168,0.1433724429872301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,256,0.0034924443397257063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,3584,0.007978666987684038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,1024,0.002874666617976295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,2560,0.016190222567982145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,768,0.004914666629499859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,8192,0.05473866727617052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,2560,0.007251555720965068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,1536,0.02200800014866723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,1536,0.007840000092983246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,65536,0.10581955644819473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,2048,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,768,0.0035742223262786865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,128,0.019680889116393197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,5120,0.012133333418104382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,16384,0.030039111773173015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,3584,0.024693333440356787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,768,0.010685332947307162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,1024,0.006250666661394968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,768,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,10240,0.036333332459131874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,8192,0.024671110841963027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,5120,0.018924444913864136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,5120,0.0137155552705129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,10240,0.012332444389661154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,2048,0.04564177658822802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,4096,0.10966933435863918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,512,0.01110666659143236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,2560,0.007284444239404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,16384,0.04658044377962748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,16384,0.18996711572011313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,512,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,768,0.02463022205564711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,3072,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,32,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,5120,0.009371555513805812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,8192,0.01033688916100396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,768,0.017177777157889474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,8192,0.007670222057236566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,1024,0.005583111196756363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,2048,0.008974221845467886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,256,0.00832622249921163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,128,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,3584,0.028487109475665625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,12288,0.02644088864326477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,1024,0.023381332556406658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,4096,0.014133334159851074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,2048,0.014107555150985718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,32,0.002152888932161861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,128,0.00629688882165485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,128,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,32,0.0018373332503769135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,8192,0.21329600281185576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,6144,0.024031110935741003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,1024,0.00554666668176651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,2048,0.008177777959240807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,16384,0.05580888854132759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,256,0.0028657778683635923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,1024,0.028927998410330877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,7168,0.026817777090602454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,10240,0.04892177714241875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,512,0.013472889032628802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,1536,0.01186222251918581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,65536,0.08571822113460964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,7168,0.012512000070677863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,8192,0.011042666931947073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,1024,0.017186666528383892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,32,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,8192,0.029494222667482164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,768,0.007102222078376346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,12288,0.28609689076741535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,3072,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,512,0.0028604444944196274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,768,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,8192,0.015486222174432544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,1536,0.018552000323931377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,3072,0.02093688812520769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,768,0.021761778328153823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,768,0.007269333634111617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,3584,0.009284444153308868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,2560,0.007602666815121968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,12288,0.04948266678386264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,16384,0.06774399677912395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,65536,0.05137688914934794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,2048,0.006304889089531369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,6144,0.04244000050756666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,7168,0.004918222212129169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,1536,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,512,0.002477333363559511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,12288,0.6720915370517306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,2048,0.006847111301289664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,8192,0.030239999294281006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,2048,0.0176862229903539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,65536,0.5011430846320258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,3584,0.05450310972001818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,32,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,5120,0.015210666590266757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,2048,0.047629333204693265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,6144,0.03536444571283128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,4096,0.010112889111042023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,7168,0.038486222426096596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,65536,0.07656621932983398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,2560,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,10240,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,65536,0.13412977589501276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,768,0.01516888870133294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,3072,0.0070222218831380205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,65536,0.2699057790968153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,2048,0.13015911314222547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,128,0.0144213338692983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,8192,0.13856800397237143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,256,0.0027875554644399216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,12288,0.01789155602455139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,3072,0.02165066699186961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,512,0.0031804444475306403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,65536,0.9910329182942709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,7168,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,4096,0.011011555790901184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,6144,0.03254400028122796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,512,0.005239999956554837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,512,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,5120,0.03435555431577895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,512,0.005642666584915585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,5120,0.01034222212102678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,10240,0.03352533446417914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,6144,0.009296889106432596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,768,0.005647111270162795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,12288,0.0518817769156562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,1536,0.003539555602603488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,128,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,128,0.002136888810329967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,12288,0.016505777835845947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,6144,0.38075465626186794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,2560,0.010253333383136326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,6144,0.009194666312800514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,65536,0.17349510722690156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,3072,0.015107555521859063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,12288,0.01648622254530589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,64,0.0032319999817344877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,2048,0.03537333342764113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,7168,0.02458311120669047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,2560,0.006666666517655055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,2560,0.015434667468070984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,65536,0.14512266053093806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,10240,0.02426044477356805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,1024,0.01761511133776771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,16384,0.1507928901248508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,2048,0.006928000185224745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,256,0.014081777797804939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,16384,0.06672355863783094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,768,0.005304000029961268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,5120,0.011366222467687396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,10240,0.017837334010336135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,8192,0.06263199779722425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,128,0.0192275560564465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,65536,0.07308889097637601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,5120,0.24391643206278482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,128,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,4096,0.007039111521508958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,2560,0.014085332552591959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,5120,0.013632888595263163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,16384,0.1687075561947293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,512,0.005186666631036335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,2048,0.015511999527613321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,10240,0.018943111101786297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,2048,0.0042160000238153665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,12288,0.024992000725534227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,512,0.02439911166826884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,16384,0.06652177704705133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,5120,0.00833511104186376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,3072,0.035938666926489934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,8192,0.02975822157329983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,256,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,512,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,512,0.006609777609507243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,16384,0.035353776481416486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,65536,0.011370666325092316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,1536,0.034952001439200506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,512,0.0024577778660588795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,4096,0.0075804442167282104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,768,0.003170666595300039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,16384,0.030268443955315485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,5120,0.008343111309740279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,768,0.0031715554909573663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,1024,0.018938667244381376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,512,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,65536,0.54239469104343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,3072,0.016368000043763053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,12288,0.27495911386277944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,4096,0.025407110651334126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,3072,0.007992000215583378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,4096,0.024492444263564214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,65536,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,65536,0.008982222113344405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,2560,0.012464000119103326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,256,0.0057760001056724125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,128,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,256,0.014829332629839579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,512,0.013461333182122974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,1536,0.302839994430542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,7168,0.009813333551088968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,512,0.014488000008794995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,128,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,16384,0.016858667135238647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,8192,0.01776977711253696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,3584,0.0793244441350301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,6144,0.1700915495554606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,65536,0.18887021806504992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,768,0.008672888908121321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,3584,0.005942222144868638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,4096,0.007986666427718269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,1024,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,32,0.0021840000732077495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,256,0.004934222333961063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,6144,0.030179556873109605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,256,0.0035777779089079965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,3584,0.018206222189797294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,8192,0.017380444539917838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,16384,0.014771555860837301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,6144,0.008960888617568547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,2560,0.04719110992219713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,32,0.001812444378932317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,1536,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,3584,0.15218488375345865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,6144,0.009299555586444007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,5120,0.09509511126412286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,12288,0.022625777456495497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,8192,0.019689778486887615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,2560,0.0069155556460221606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,12288,0.016593777471118502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,256,0.00544177782204416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,512,0.003175111073586676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,3072,0.018527110417683918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,768,0.014769777655601501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,12288,0.015104888214005364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,1536,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,8192,0.01947733263174693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,256,0.01615111033121745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,5120,0.012788444757461548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,65536,0.1372364494535658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,6144,0.04668177829848396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,12288,0.04983377787801954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,10240,0.0383840004603068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,32,0.011019555230935415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,7168,0.06111288732952542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,3072,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,6144,0.15317155255211726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,5120,0.018929777873886954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,64,0.006578666468461354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,768,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,1536,0.006238222122192383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,5120,0.028776887390348647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,512,0.0028364445186323593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,3072,0.012412444584899478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,16384,0.02404088940885332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,65536,0.5937697622511122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,3584,0.006588444527652528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,1536,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,2048,0.005168000029193031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,128,0.014426666829321118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,256,0.02185955478085412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,3584,0.023006222314304773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,6144,0.03529599971241421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,7168,0.009678222239017487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,3584,0.008833777573373582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,12288,0.38901599248250324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,3584,0.021583999196688335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,512,0.006062222023804982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,32,0.005951111101441913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,4096,0.08205422427919176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,1024,0.0388302206993103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,7168,0.13246488571166992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,16384,0.03824533356560601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,4096,0.03388977713055081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,7168,0.01721333298418257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,7168,0.01791822248035007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,3072,0.008674666285514832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,4096,0.011013333168294696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,4096,0.04168533285458883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,1024,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,16384,0.23606843418545195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,6144,0.09613422552744548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,6144,0.004927111168702443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,2048,0.014456000592973499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,2560,0.0192239996459749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,256,0.005007110950019625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,3584,0.05081955591837565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,8192,0.010334221853150262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,16384,0.06466399961047702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,16384,0.015098666151364645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,8192,0.10299822356965806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,3584,0.0240666667620341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,1024,0.0075911109646161394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,3072,0.009676444033781687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,12288,0.05787999762429131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,64,0.006284444282452266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,256,0.018211555149820115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,128,0.004903111192915174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,2048,0.028304888142479792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,3584,0.04199111130502489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,256,0.013822221921549903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,512,0.00281866660548581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,512,0.006945778098371293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,512,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,2560,0.021589333812395733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,512,0.006229333165619109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,1536,0.0038497779104444715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,7168,0.012899555265903473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,256,0.016304888659053378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,6144,0.009007111191749573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,1536,0.006224888894293044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,16384,0.06842311223347981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,3072,0.017852443787786696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,512,0.005210666606823604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,768,0.01756444407833947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,512,0.014891554911931356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,768,0.007670222057236566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,512,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,2048,0.073999113506741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,1024,0.0032008888406885993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,3584,0.011050666371981302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,8192,0.022993778189023335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,256,0.01739911072784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,4096,0.14880888991885716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,128,0.024432000186708238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,2560,0.04089777668317159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,4096,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,2560,0.008033778104517195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,2048,0.0064995553758409284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,128,0.0053013331360287136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,128,0.005863111052248213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,2560,0.008707555631796518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,1024,0.005577777822812398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,512,0.0028168888141711554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,5120,0.34876177046034074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,7168,0.011672889192899069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,10240,0.023036445180575054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,768,0.0028577778074476454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,12288,0.027134221461084154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,1024,0.008700444466537898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,3584,0.007290666302045186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,12288,0.17845243877834746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,768,0.017486222916179232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,64,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,1024,0.004861333303981357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,1024,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,256,0.015477332803938123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,2048,0.022287999590237934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,8192,0.012531555361217923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,256,0.016174222032229107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,3584,0.0181404451529185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,5120,0.017863111363516915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,4096,0.0105253333846728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,2048,0.008286221987671321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,10240,0.01792266633775499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,1536,0.004297777596447202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,8192,0.00543200017677413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,768,0.0032275555034478507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,12288,0.027247111002604168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,12288,0.019917334119478863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,10240,0.011362666885058085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,128,0.009022222624884712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,7168,0.01886666648917728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,512,0.006303999986913469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,12288,0.18303110864427355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,8192,0.028947555356555518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,512,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,3584,0.013946667313575745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,16384,0.13177066379123265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,16384,0.05196355448828804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,4096,0.009319999979601966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,8192,0.05784355269538032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,65536,0.06491111384497748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,4096,0.08190666966968112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,7168,0.00619911112719112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,1024,0.029472000069088403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,1024,0.0028693332440323303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,4096,0.019464888506465487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,128,0.002817777709828483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,256,0.01475288967291514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,6144,0.028095109595192805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,768,0.004889777965015835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,256,0.0025457777082920074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,12288,0.05166044500139025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,2560,0.00980355590581894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,32,0.005580444302823808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,7168,0.014186667071448432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,256,0.01661777827474806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,12288,0.020258666740523446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,6144,0.012741333080662621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,3072,0.014863999353514777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,7168,0.011413333316644033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,128,0.0028124445428450904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,12288,0.041788445578681103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,7168,0.0336097776889801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,128,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,1536,0.021669333179791767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,6144,0.01163999984661738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,7168,0.021520000365045335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,1536,0.006237333433495627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,32,0.011757333245542316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,10240,0.030774222479926214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,6144,0.009353777600659264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,10240,0.018586667047606576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,8192,0.1784400012758043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,2048,0.015509333875444202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,16384,0.02271822260485755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,4096,1.4876968595716689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,1536,0.007344000041484833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,6144,0.03706222110324436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,12288,0.043134222428003945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,16384,0.027626666757795546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,12288,0.017833777599864535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,32,0.0018079999006456798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,1536,0.0036755556033717263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,128,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,5120,0.008984888593355814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,6144,0.02199199961291419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,1024,0.01832355558872223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,16384,0.02904799911710951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,128,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,4096,0.009311999711725447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,3584,0.022198221749729578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,64,0.024533333049880132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,5120,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,32,0.0018666666001081467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,128,0.014087110757827759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,65536,0.10275555319256252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,6144,0.011230222052998014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,256,0.0031644445326593188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,32,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,4096,0.007937777373525832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,4096,0.015817777978049386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,3584,0.0473599996831682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,64,0.007296000089910295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,12288,0.08201599783367582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,256,0.004902222090297275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,64,0.003165333428316646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,4096,0.007616889145639207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,7168,0.010033778018421596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,1536,0.007654222349325816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,32,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,5120,0.04862933357556661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,8192,0.024567999773555334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,3072,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,256,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,16384,1.4474035898844402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,5120,0.010011555420027839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,128,0.0035386664999855887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,256,0.018598222070270114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,3072,0.03812711106406318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,5120,0.04119911127620273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,7168,0.010407110883129967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,10240,0.02371999952528212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,256,0.005561777700980504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,256,0.006217777729034424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,16384,0.024353777368863423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,16384,0.01516355574131012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,2048,0.006272000157170826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,64,0.018264888061417472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,128,0.002206222257680363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,1536,0.02514044443766276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,6144,0.016089777151743572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,10240,0.015215999550289579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,768,0.0028897778441508612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,1024,0.04256088866127861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,6144,0.11468266116248237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,2048,0.004290666845109728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,1024,0.02288533416059282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,768,0.0028737777223189673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,64,0.0021093332519133887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,128,0.02033955521053738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,768,0.005943111247486538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,10240,0.04559200008710226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,256,0.011777777638700275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,16384,0.07840089003245036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,65536,0.008640888664457533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,2048,0.01906399925549825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,8192,0.012290666500727335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,1536,0.009340444372759925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,768,0.007283555964628856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,768,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,2048,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,12288,0.02882222334543864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,1024,0.012727110750145383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,3072,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,65536,0.10170311397976345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,10240,0.016072000066439312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,4096,0.6518221961127387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,8192,0.03391289048724704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,16384,0.01585422290696038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,7168,0.014857777290874057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,4096,0.02053333322207133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,2048,0.026828444666332666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,12288,0.01802399920092689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,12288,0.0792142219013638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,128,0.012736000120639801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,8192,0.053413334820005626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,5120,0.12298133638170029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,6144,0.019097778532240126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,512,0.006578666468461354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,2560,0.008262222011884054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,3584,0.025750221477614507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,12288,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,2560,0.01447111037042406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,128,0.0028408887899584244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,6144,0.015799999237060547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,1024,0.02644799980852339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,3584,0.03864977757136027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,12288,0.050832887490590416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,2048,0.029725333054860432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,65536,0.07992977566189237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,16384,0.06537866592407227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,3072,0.010038221875826517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,65536,0.1896737813949585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,256,0.00216977774269051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,768,0.0028195555011431375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,64,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,7168,0.013260444833172692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,12288,0.013771555489963956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,16384,0.013438222308953604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,32,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,64,0.0017973333597183228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,3584,0.010048000348938836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,256,0.010059555371602377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,12288,0.0277528895272149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,8192,0.023312888211674158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,6144,0.009024889104896123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,5120,0.10682044426600139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,512,0.011078221930397881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,3072,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,3584,0.010021333065297868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,512,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,3072,0.01550044450494978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,1024,0.020298666424221463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,256,0.016156444946924847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,2560,0.021949332621362474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,2560,0.006611555400821898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,3072,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,8192,0.010223111344708337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,3584,0.06540355417463514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,7168,0.029567999972237483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,2560,0.06521688567267524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,4096,0.039840890301598444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,32,0.0018417777286635505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,65536,0.09496977594163682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,3584,0.007987555530336168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,10240,0.017768000562985737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,1024,0.005879111174080107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,32,0.0021359999146726397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,2048,0.006468444648716185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,2048,0.015526221858130561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,512,0.005199999858935674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,32,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,8192,0.01588088936275906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,1024,0.015133332875039844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,1024,0.003547555456558863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,1024,0.005690666536490123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,1536,0.0034924443397257063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,2560,0.009593778186374241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,2048,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,6144,0.03417244553565979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,768,0.009191999832789103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,6144,0.04518489042917887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,7168,0.020587555236286588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,4096,0.012730666332774691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,7168,0.02256711158487532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,6144,0.012109333442317115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,1024,0.017256889078352187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,4096,0.011899555722872416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,4096,0.007619555625650618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,512,0.006903111106819577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,3072,0.01185155577129788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,7168,0.14459733168284097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,6144,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,16384,0.018904889623324077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,5120,0.027826666831970215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,1024,0.06052177482181125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,8192,0.010764444039927589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,3072,0.016547555724779766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,128,0.008977777428097194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,16384,0.06150755617353651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,3584,0.010447999669445885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,8192,0.04966310991181267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,256,0.03290844294759963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,2048,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,128,0.0052328887912962176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,64,0.001840888833006223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,32,0.00527022240890397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,3072,0.007301333049933116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,6144,0.014793778459231058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,256,0.004981333182917701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,12288,0.0964026649792989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,2560,0.01514755520555708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,16384,0.0243857784403695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,2560,0.007622222105662028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,7168,0.011920000116030375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,256,0.006953777538405524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,7168,0.04424355427424113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,8192,0.00926755534278022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,1024,0.006934222247865465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,10240,0.01791911158296797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,7168,0.038005332152048744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,16384,0.01682844426896837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,64,0.0018088887963030075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,256,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,65536,0.09924266735712688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,2560,0.23549956745571562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,8192,3.062381320529514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,768,0.01923377811908722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,16384,0.022312889496485393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,10240,0.027070222629441157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,8192,0.020340444313155282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,64,0.008408889174461365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,10240,0.02179466684659322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,3584,0.02239111065864563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,128,0.007592889169851939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,10240,0.009192888935407003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,65536,0.14737155702379015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,768,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,12288,0.024381332927280005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,5120,0.011066666907734342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,2560,0.00790666706032223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,4096,0.009017777939637503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,6144,0.017588444881969027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,16384,0.020676443974177044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,2560,0.007967111137178209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,768,0.005381333331267039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,7168,0.04948355423079597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,3072,0.009374221993817223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,32,0.001957333336273829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,5120,0.12745243973202175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,3584,0.01920622256067064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,1536,0.005952888892756567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,4096,0.006568000134494569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,64,0.0022177776942650476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,8192,0.12387111451890732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,64,0.001828444500764211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,8192,0.018086221483018663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,3584,0.011454222102959951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,12288,0.03690133492151896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,4096,0.011191999746693505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,3072,0.010703110860453712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,4096,0.007990222010347579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,2560,0.011095999843544431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,32,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,768,0.0028719999310043124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,6144,0.015531554818153381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,3584,0.023912888434198167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,64,0.006583111153708563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,1024,0.00564444437623024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,3584,0.004899555610285865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,3584,0.010968888799349466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,1536,0.03239288926124573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,768,0.00867377801073922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,32,0.0053297777970631914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,16384,0.09460355838139851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,16384,0.04343200061056349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,1024,0.006303999986913469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,3584,0.010279999838935005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,2560,0.011744000017642975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,7168,0.12669511636098227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,2048,0.006599999964237213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,12288,0.3951546616024441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,3584,0.007335111498832703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,65536,0.09779910908804999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,256,0.012430222498046027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,2048,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,65536,0.057702223459879555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,128,0.006595555692911148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,65536,0.20424444145626494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,12288,0.08282666736178927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,2048,0.017687999539905123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,4096,0.007968000239796108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,65536,0.15520800484551325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,7168,0.04899644520547655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,16384,0.022967111733224656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,3072,0.00795999997191959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,7168,0.011923555698659686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,768,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,12288,0.02527644402450985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,3072,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,7168,0.057368887795342334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,256,0.007995555798212687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,4096,0.045647998650868736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,8192,0.058407107988993325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,7168,0.00925866680012809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,6144,0.03783644570244683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,32,0.003863999827040566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,7168,0.10232977734671699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,1024,0.004567111117972268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,3072,0.11223822169833714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,512,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,1024,0.010271111296282874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,4096,0.02240266733699375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,3584,0.004511111312442356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,4096,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,6144,0.009677333136399588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,8192,0.027142221728960674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,7168,2.41412353515625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,64,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,10240,0.12967644797431097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,10240,0.032300445768568255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,10240,0.16335822476281062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,5120,0.009340444372759925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,768,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,16384,0.03573333223660787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,16384,0.09464177820417617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,128,0.014841778410805596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,512,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,2560,0.006601777755551868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,1536,0.01779911087618934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,6144,0.0118222220076455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,8192,0.030538668235143025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,3072,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,3072,0.005514666852023866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,4096,0.14852355586157903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,65536,0.1894728872511122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,8192,0.6084622277153863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,2048,0.006552000012662675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,65536,0.1790693336062961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,12288,0.009887110855844285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,256,0.006619555668698416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,32,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,8192,0.05938488907284207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,3072,0.04685066805945503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,64,0.015531554818153381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,10240,0.022335999541812476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,65536,0.22822133700052896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,2560,0.08848266469107734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,1536,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,10240,0.2500648763444689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,5120,0.014834667245546976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,128,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,6144,0.05956000089645386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,3584,0.007292444507280986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,128,0.002483555633160803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,32,0.005396444350481033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,2560,0.2781413396199544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,1024,0.005883555445406172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,65536,0.15042044056786433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,768,0.002528888897763358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,4096,0.007655999726719326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,7168,0.00797333319981893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,10240,0.035953776703940496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,10240,0.018439110782411363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,3072,0.007264888948864407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,3072,0.008923555413881937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,6144,0.03064089020093282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,32,0.0035253332720862497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,2560,0.012544000314341651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,32,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,512,0.028875556257035997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,128,0.013448889056841532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,16384,0.03571555680698819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,7168,0.029046220911873713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,16384,0.022806222240130108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,12288,0.04522044459978739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,10240,0.02661688956949446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,1024,0.006255110932721033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,2560,0.009188444250159794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,768,0.007610666255156199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,512,0.00692622239391009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,4096,0.017072889539930556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,8192,0.08822933170530532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,128,0.013420444395807056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,64,0.0064488889442549805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,3584,0.00850755555762185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,2048,0.006666666517655055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,4096,0.010209778116808997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,1024,0.01998044384850396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,5120,0.08353510830137466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,128,0.01481066644191742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,2048,0.0940328902668423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,65536,0.0953528881072998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,4096,0.007639110916190677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,2560,0.006615110983451207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,2048,0.00832622249921163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,512,0.010964444114102257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,16384,0.027820444769329492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,8192,0.024412444896168176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,512,0.008316444853941599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,8192,0.02093422247303857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,8192,0.03852355480194092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,16384,0.023860444625218708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,2560,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,2048,0.007259555160999298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,5120,0.20717599656846789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,5120,0.005693333430422678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,768,0.015491555134455362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,2048,0.006262222097979651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,1536,0.06927733288870917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,12288,0.04936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,3584,0.018234666850831773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,4096,0.009583999713261923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,3072,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,1024,0.005616000129116907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,1024,0.006257777826653586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,4096,0.011373333632946014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,3584,0.011287111375066968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,4096,0.025420443879233465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,2048,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,2560,0.09626933601167466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,5120,0.007661333514584436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,768,0.003208000005947219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,16384,0.034324444002575345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,2560,0.0069120000633928515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,1536,0.005882666756709416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,1536,0.005917333480384614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,3072,0.0075697774688402815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,128,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,12288,0.11615999539693196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,16384,0.07124266359541151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,2048,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,7168,0.15178133381737605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,16384,0.015088000231319003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,7168,0.027169777287377253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,8192,0.03404622276624044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,7168,0.010094222095277574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,5120,0.01242844429281023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,3072,0.007300444775157505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,2560,0.043064888980653554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,3584,0.01402933398882548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,3584,0.020022221737437777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,1536,0.011705777711338468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,256,0.005873777800136142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,128,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,1536,0.01239911135700014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,16384,0.023032888770103455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,5120,0.0224924451775021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,10240,0.04074222180578444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,65536,0.07959733406702678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,2048,0.10494133498933579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,6144,0.012725333372751871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,2560,0.015450666348139444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,768,0.03423377871513367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,2048,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,2560,0.8777404361300998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,4096,0.04642577634917366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,12288,0.021338666478792827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,2560,0.013439999686347114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,768,0.05278844303554959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,65536,0.22075022591484916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,768,0.017892445127169292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,32,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,3072,0.007742221984598372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,1536,0.00850399997499254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,4096,0.011389333340856763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,3584,0.010066666536860997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,2048,0.012368888490729861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,7168,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,4096,0.018805333309703402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,10240,0.025399999486075506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,768,0.0025333333760499954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,3072,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,256,0.0063573333124319715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,16384,0.0358497765329149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,2048,0.020561777883105807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,256,0.005691555639108022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,512,0.006007110906971826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,128,0.010055999788973067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,4096,0.00811822215716044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,64,0.005335111171007156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,5120,0.025363556212849084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,768,0.003584889074166616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,64,0.005792888916201062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,64,0.0039013334446483185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,1536,0.015120888749758402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,65536,0.09089155329598321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,1024,0.01275111072593265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,512,0.019612444771660697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,2560,0.0069120000633928515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,2048,0.006577777779764599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,8192,0.013697778185208639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,5120,0.02346399923165639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,768,0.006719999843173557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,4096,0.04755022128423055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,256,0.014149333039919535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,1536,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,5120,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,768,0.02067466742462582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,8192,0.046047998799218066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,2048,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,12288,0.05066844489839342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,5120,0.008624888956546783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,256,0.015876443849669564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,8192,0.2173759937286377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,1536,0.005966222120655908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,128,0.0035706667436493766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,12288,0.17052622636159262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,1024,0.013415999710559845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,3584,0.007601777712504069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,8192,0.02986488739649455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,12288,0.018009778526094224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,8192,0.03506133291456435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,7168,0.009690666364298927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,5120,0.022650667362742957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,128,0.005247999810510212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,65536,0.44083378050062394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,2048,0.031572444571389094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,2560,0.004960000101062986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,4096,0.003903111235962974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,6144,0.004882666799757215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,2048,0.006765333314736684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,4096,0.009033777647548253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,64,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,64,0.007343110938866933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,5120,0.01885955532391866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,6144,0.03167377909024557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,768,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,65536,0.24300800429450142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,5120,0.12105511294470893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,64,0.006583111153708563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,4096,0.0075911109646161394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,6144,0.0123831108212471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,64,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,2560,0.016170667277442086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,2048,0.006958222223652734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,512,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,4096,0.011439999772442712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,1024,0.009444444543785518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,512,0.03310311171743605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,7168,0.023341332872708637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,10240,0.035572446054882474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,4096,0.015863999724388123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,128,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,10240,0.021317332983016968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,4096,0.016809778081046212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,768,0.029920889271630183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,7168,0.01573244399494595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,256,0.01444533301724328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,7168,0.008988444175985124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,10240,0.022083555658658344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,7168,0.022023111581802368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,512,0.012806221842765808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,1024,0.015126221709781222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,10240,0.028466665082507666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,768,0.01127199994193183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,3072,0.2709280120001899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,32,0.003175111073586676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,3584,0.03219644559754266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,5120,0.008480889101823172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,3584,0.009995555712117089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,16384,0.026407111022207472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,4096,0.028024001253975764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,12288,0.2291173405117459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,12288,0.022239110536045496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,256,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,65536,0.078621334499783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,768,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,1536,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,6144,0.009995555712117089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,7168,0.025849777791235182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,3072,0.007264888948864407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,5120,0.00830666638082928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,12288,0.014127110441525778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,3584,0.04858933223618401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,768,0.0184897780418396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,16384,0.0684773325920105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,512,0.004905777672926585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,4096,0.020272889071040683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,10240,0.056886222627427846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,10240,0.03973777757750617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,12288,0.021903999977641638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,4096,0.004234666625658671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,256,0.01070844464831882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,64,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,6144,0.013791999883121915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,64,0.0018328889790508484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,2560,0.021976888179779053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,65536,0.07542666461732653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,16384,0.03459822138150533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,4096,0.007325333025720384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,6144,0.39867377281188965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,4096,0.007903999752468532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,256,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,12288,0.28596356179979115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,10240,0.009360888765917884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,10240,0.014706666270891825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,32,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,128,0.02130844526820713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,12288,0.0829662217034234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,768,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,2560,0.28921156459384495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,256,0.0052328887912962176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,1024,0.005945777727497949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,2048,0.05877955754597982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,1536,0.013791999883121915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,512,0.006217777729034424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,6144,0.009016888837019602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,32,0.0021928888228204516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,8192,0.03388177686267429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,256,0.015484443969196744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,2048,0.030067556434207495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,768,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,12288,0.032031112247043185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,6144,0.261445336871677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,3584,0.007276444799370236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,6144,0.02274133265018463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,2560,0.006647111227114995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,3584,0.007678222325113084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,2560,0.024715556038750544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,768,0.005881777654091517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,1024,0.012038222617573209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,128,0.024216888679398432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,12288,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,768,0.03560444381501939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,7168,0.012789333032237159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,1024,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,4096,0.007281777759393056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,4096,0.008000888758235509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,1536,0.006962666908899943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,256,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,128,0.047015110651652016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,4096,0.020206222931543987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,768,0.00291377781993813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,1024,0.02367555598417918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,1536,0.015635554989178974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,2048,0.013798221945762634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,256,0.02274311085542043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,10240,0.03528799944453769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,1024,0.006311111152172089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,12288,0.33583911259969074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,512,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,6144,0.01779022150569492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,16384,0.04261066516240438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,7168,0.015917332635985482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,256,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,12288,0.05742488967047798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,2048,0.019926221834288705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,7168,0.017893332574102614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,65536,0.07579200135336982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,768,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,2048,0.02591288917594486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,6144,0.033505777517954506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,768,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,256,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,10240,0.04526311159133911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,3584,0.08665155702167088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,3584,0.00796888851457172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,512,0.011766222616036734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,5120,0.016014221641752455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,7168,0.030277334981494482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,32,0.1679839955435859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,256,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,10240,0.27957333458794487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,4096,0.016179554992251925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,128,0.004929777648713854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,512,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,3584,0.015856888559129503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,256,0.0034959999223550162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,128,0.014504000544548035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,5120,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,3584,0.05236711104710897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,7168,0.017708443933063082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,1536,0.005938666562239329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,3584,0.023136888941129048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,3072,0.01350311107105679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,5120,0.0546595553557078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,64,0.0022257777551809945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,512,0.005956444475385878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,12288,0.012240000069141388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,2560,0.3658497863345676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,128,0.004864000197913912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,768,0.016912889149453905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,7168,0.054139554500579834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,10240,0.10643911361694336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,3584,0.015132443772421943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,12288,0.03683999843067593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,32,0.024597333537207708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,128,0.002132444539003902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,5120,0.045258667733934194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,3072,0.06830310821533203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,768,0.002492444382773505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,7168,0.038959999879201256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,3584,0.007616000043021307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,1536,0.0035342222286595237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,3072,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,8192,0.010976889067225985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,10240,0.015448000695970325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,768,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,3072,0.014560000763999091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,12288,0.034556445148256086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,5120,0.0083137775460879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,12288,0.017501332693629794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,10240,0.17743821938832602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,6144,0.02771200074089898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,2560,0.006588444527652528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,256,0.004885333279768626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,1536,0.010306666294733683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,1536,0.006217777729034424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,3584,0.023794665932655334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,3072,0.007595555649863348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,768,0.03506933318244086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,1024,0.051983998881445996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,10240,0.03957155678007338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,16384,0.06978578037685819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,64,0.0053048887186580235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,65536,0.3286657863193088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,1024,0.0058782220714622075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,5120,0.03731377919514974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,512,0.0028577778074476454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,12288,0.05841155846913656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,4096,0.004496888981925117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,16384,0.017161778277821012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,3072,0.020322667227851022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,16384,0.04351111253102621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,1536,0.01688799924320645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,4096,0.07768444220225017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,65536,0.9760061899820963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,6144,0.016950221525298226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,7168,0.011376889215575324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,1024,0.007257777783605788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,256,0.002171555534005165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,512,0.03399822115898132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,64,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,12288,0.02756711178355747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,2560,0.02068355513943566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,2560,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,12288,0.05759555763668484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,12288,0.013564444250530668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,6144,0.013408888545301227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,2560,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,3584,0.007653333246707916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,32,0.006924444602595435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,512,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,1536,0.032687111033333674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,5120,0.013400000002649097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,5120,0.01644088824590047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,3584,0.01797777745458815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,1024,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,3072,0.02091377807988061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,3072,0.01649599936273363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,512,0.01888088881969452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,2048,0.010159999959998662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,32,0.002537777854336633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,64,0.002151111140847206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,64,0.001850666685236825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,1536,0.019662222928471036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,1024,0.025660443637106154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,128,0.0021315556433465746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,4096,0.024062222904629175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,7168,0.01533422205183241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,1024,0.017216889394654166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,128,0.005481777919663324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,3072,0.01035377797153261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,16384,0.06352000104056464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,10240,0.012131555212868584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,128,0.00565155554148886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,5120,0.03368355499373542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,4096,0.017532444662517972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,6144,0.016525333126386006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,3584,0.005895111295912001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,7168,0.040469331873787776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,2560,0.007481777833567725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,65536,0.1803039974636502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,12288,0.37393689155578613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,512,0.005320888840489917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,6144,0.0233297778500451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,6144,0.02198311189810435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,1024,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,128,0.015447111593352424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,65536,0.06749422020382352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,4096,0.019305777218606737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,10240,0.813071992662218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,1024,0.02131822208563487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,6144,0.02959733208020528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,1024,0.006673777682913675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,512,0.01888177792231242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,3072,0.012949333422713809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,768,0.03997600078582764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,5120,0.011691555380821228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,3584,0.02808977829085456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,7168,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,1536,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,10240,0.07371289200252958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,8192,0.08901866939332749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,8192,0.028443555037180584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,4096,0.007667555577225155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,1536,0.006660444455014334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,6144,0.010704889065689512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,1024,0.05490577883190579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,65536,0.07594755623075697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,12288,0.09457688861423069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,1024,0.007988444632954067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,768,0.007289778027269576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,3072,0.011433777709801992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,4096,0.016864000095261466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,2048,0.014134221606784396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,10240,0.02404977712366316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,6144,0.027191110783153113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,16384,0.632080872853597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,4096,0.01797777745458815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,6144,0.029143999020258587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,128,0.014142221874660917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,128,0.002489777695801523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,256,0.013136000268989138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,3072,0.006970666348934174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,65536,0.10920533206727769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,4096,0.010413332945770687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,8192,0.023316444622145757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,6144,0.005993777679072485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,5120,0.01744888888465034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,3584,0.017303110824690927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,8192,0.009741333623727163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,2048,0.00831200016869439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,12288,0.018218666315078735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,256,0.003544888976547453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,8192,0.028370665179358587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,4096,0.01205244412024816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,256,0.01716800034046173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,8192,0.015156444576051501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,2560,0.012882666455374824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,2560,0.006871999965773688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,12288,0.02330488959948222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,512,0.009699555734793345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,768,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,6144,0.008959111240175035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,7168,0.016134222348531086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,128,0.003175111073586676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,256,0.01444533301724328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,768,0.004564444637960858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,2560,0.03340355555216471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,2560,0.02067377832200792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,768,0.0038684445122877755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,16384,0.03756711218092177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,128,0.014152000347773233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,65536,0.5990222295125326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,256,0.003547555456558863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,2048,0.01854222185081906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,2048,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,3584,0.039108445247014366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,256,0.014442667365074158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,7168,0.12053688367207845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,2560,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,8192,0.027752000424597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,5120,0.059439109431372754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,64,0.0017955555684036678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,65536,0.1566284497578939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,512,0.006803555621041193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,12288,0.03579466541608175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,512,0.025615110993385315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,10240,0.14244888888465032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,3584,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,16384,0.017872000734011333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,768,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,768,0.014469333820872836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,65536,0.24742666880289713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,1536,0.007928888830873702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,512,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,8192,0.6061840057373047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,12288,0.3434266779157851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,8192,0.024199111594094172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,65536,0.06083910995059543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,12288,0.013850666582584381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,3072,0.010507555471526252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,4096,0.010568888651000129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,16384,0.018612444400787354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,10240,0.014306666122542487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,3584,0.00814844419558843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,16384,0.06519377893871732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,4096,0.020290666156344943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,2560,0.0196079992585712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,4096,0.007313778003056844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,64,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,7168,0.004250666747490565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,16384,0.04970755510860019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,1024,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,6144,0.012652444342772165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,1536,0.00665511108107037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,7168,0.010856000085671743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,1024,0.013747555514176687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,1536,0.047261334127850003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,6144,0.0340595543384552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,16384,0.23373243543836805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,64,0.002187555655837059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,65536,0.06845155689451429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,16384,0.11527910497453477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,65536,0.3991857899559869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,512,0.005535111245181825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,256,0.026735110415352717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,7168,0.010011555420027839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,7168,0.013462222284740873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,1536,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,256,0.005924444645643234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,768,0.00693866651919153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,1024,0.01684533390733931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,2560,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,1536,0.01015377789735794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,4096,0.03465955456097921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,1024,0.06136622031529745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,128,0.015503111812803479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,12288,0.5034097565544976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,3072,0.008325333396593729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,8192,0.015535111228624979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,2560,0.007164444360468123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,3072,0.008650666309727563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,4096,0.004581333448489507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,6144,0.02163377735349867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,128,0.006592000110281839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,512,0.0031519999934567344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,7168,0.012746666868527731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,16384,0.014654222461912366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,1536,0.015495111544926962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,128,0.00518133325709237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,5120,0.0042239998777707415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,8192,0.011408000356621213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,8192,0.46799288855658633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,1024,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,2560,0.010689777632554373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,512,0.0038977778620190094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,5120,0.017709333035680983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,256,0.004940444396601783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,12288,0.042280889219707914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,2560,0.007744889292452071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,1536,0.004913777940803104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,10240,0.054766221178902515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,64,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,4096,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,4096,0.007445333732499017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,5120,0.04394488864474826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,6144,0.011018666956159802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,128,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,512,0.006618666566080517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,4096,0.014188443620999655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,7168,0.036882665422227644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,256,0.007621333003044128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,65536,0.1434897714191013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,1536,0.0063182223174307085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,1536,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,8192,0.02032177812523312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,256,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,65536,0.0714871088663737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,4096,0.01847466660870446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,32,0.006015111174848344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,5120,0.004882666799757215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,4096,0.007985778152942657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,65536,0.08066488636864556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,2560,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,1024,0.004907555464241239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,256,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,2560,0.01276177747382058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,6144,0.028889778587553237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,65536,0.07455644342634413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,12288,0.02784088916248745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,3072,0.06088978052139282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,65536,0.08174310790167914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,512,0.006589333216349284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,4096,0.0995840032895406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,6144,0.07830399937099881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,768,0.00282222218811512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,1536,0.015793777174419828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,7168,0.010040889183680216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,2048,0.007992888490358988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,512,0.0026311110705137253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,256,0.0022888889329300984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,16384,0.015104888214005364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,3584,0.008285333712895712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,2048,0.02068000038464864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,12288,0.018810666269726224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,128,0.006214222146405115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,6144,0.034303112162484065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,128,0.018408000469207764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,256,0.004923555586073133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,12288,0.04029777646064758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,65536,3.436185836791992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,2560,0.03176000052028232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,7168,0.015627556376987033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,16384,0.04424088862207201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,4096,0.00831911133395301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,12288,0.014090667168299357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,3584,0.01475644442770216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,768,0.03430577781465318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,2048,0.00833777752187517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,768,0.015993777248594496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,256,0.006913777854707506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,128,0.004918222212129169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,10240,0.017888888716697693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,512,0.008967111508051554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,6144,0.1432924403084649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,512,0.005201777650250329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,10240,0.03065955638885498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,10240,0.017229333519935608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,12288,0.035417778624428645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,128,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,10240,0.025563556287023757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,512,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,1024,0.005956444475385878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,256,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,65536,0.2227191130320231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,10240,0.019334221879641216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,3072,0.018607111440764535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,3584,0.011428444749779172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,1536,0.006299555715587404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,12288,0.043581333425309926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,64,0.0021093332519133887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,5120,0.017121778594122995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,768,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,16384,0.019356444478034973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,6144,0.012411555482281579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,1024,0.006956444846259222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,128,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,512,0.02553155521551768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,6144,0.013239999612172445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,10240,0.008004444340864817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,512,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,12288,0.04142489035924276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,65536,0.006959111326270633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,768,0.006224888894293044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,12288,0.02165511084927453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,5120,0.023472888602150813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,128,0.006324444380071428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,3584,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,3584,0.01071200023094813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,4096,0.01756088932355245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,3584,0.007459555235173967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,32,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,16384,0.034095999267366194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,256,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,768,0.007242666350470648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,64,0.0018702221827374564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,16384,0.05418311225043403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,768,0.00664533343580034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,128,0.002144888871245914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,256,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,6144,0.01128533316983117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,128,0.005976888868543837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,7168,0.03446400165557861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,4096,0.020473778247833252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,65536,0.20179200172424316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,1536,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,16384,0.025409777959187824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,12288,0.02235466738541921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,2560,0.007144889069928064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,3584,0.029927998781204224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,4096,0.022982221510675218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,3072,0.05906933546066284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,1536,0.14377066824171278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,12288,0.012784889174832238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,10240,0.017644445101420086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,12288,0.4839208920796712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,4096,0.01146133326821857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,1024,0.009945777555306753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,64,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,2048,0.02332533399264018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,5120,0.011759111450778114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,10240,0.16069689061906603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,16384,0.025758221745491028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,6144,0.05003289050526089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,7168,0.04356000158521864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,6144,0.0131217779384719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,12288,0.04138311081462436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,8192,0.02254933284388648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,768,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,6144,0.021141333712471858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,1024,0.032877332634396024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,12288,0.03338222371207343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,768,0.014764444695578681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,3584,0.02392533255947961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,7168,0.011767999993430244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,1024,0.01720533271630605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,64,0.005607999861240387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,2560,0.11981244881947835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,10240,0.027093332674768236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,7168,0.00979200005531311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,64,0.0034862222770849862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,65536,0.09186666541629368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,1536,0.2888408766852485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,8192,0.043325334787368774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,1024,0.0031831111345026228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,5120,0.010458666417333813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,5120,0.017788444956143696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,768,0.016916443904240925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,7168,0.015475556254386902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,2048,0.01390044391155243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,256,0.004923555586073133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,12288,0.013761777844693927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,512,0.012560000022252401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,10240,0.014881778094503613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,512,0.005886222339338726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,1024,0.007350222104125553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,4096,0.017246221502621967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,1024,0.0031555555760860443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,12288,0.04873422119352552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,5120,0.07540177636676364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,6144,0.01240533341964086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,12288,0.013139555851618448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,768,0.026103999879625108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,8192,0.011548444628715515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,2560,0.015695111619101632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,5120,0.01275111072593265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,1536,0.006245333287451003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,3584,0.031168888012568157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,256,0.024884444144037034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,7168,0.009367999931176504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,16384,0.09812621937857734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,7168,0.014166222678290473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,3584,0.01386666629049513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,8192,0.012443555725945367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,2560,0.8115582466125488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,3072,0.0365555551317003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,4096,0.014398222168286642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,64,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,5120,0.013471110827393003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,4096,0.025046222739749487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,64,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,768,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,1536,0.005506666584147348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,12288,0.10836355553732978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,8192,0.028548445966508653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,6144,0.02920444475279914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,3584,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,7168,0.00998044427898195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,6144,0.023338667220539514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,3584,0.010786666638321347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,4096,0.005935110979610019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,4096,0.00792622235086229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,2560,0.008935111264387766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,1024,0.01686755485004849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,256,0.020615110794703167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,2048,0.013117333253224691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,3072,0.00701422244310379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,1024,0.01583377851380242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,3072,0.011032000184059143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,6144,0.03150399857097202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,128,0.01479555500878228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,2048,0.025068443682458665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,5120,0.014151111245155334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,5120,0.03740977909829881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,2048,0.00924800005224016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,16384,0.03750755720668369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,64,0.0018088887963030075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,7168,0.2005537748336792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,65536,0.07253066698710124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,16384,0.06557066573037042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,7168,0.010654222634103564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,2560,0.005655999812814925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,1024,0.0042462220622433555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,65536,0.09649244281980728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,12288,0.01752800080511305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,256,0.0022071111533376905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,4096,0.007931555310885111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,4096,0.012413333687517377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,12288,0.028441776831944782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,4096,0.01536266671286689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,512,0.0034577778230110803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,3584,0.014854222536087036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,3584,0.007656888829337225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,3072,0.011015999648306104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,65536,0.8103351063198514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,4096,0.03709866603215536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,2560,0.009726222190592024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,64,0.017125333348910015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,4096,0.014147554834683737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,7168,0.008808888494968414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,64,0.002030222159292963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,7168,0.009763555394278632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,5120,0.015992000699043274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,16384,0.031276444594065346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,3584,0.010764444039927589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,3072,0.008162666526105668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,1536,0.00886222223440806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,3584,0.03649866580963135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,16384,0.034332444270451866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,12288,0.3370960023668077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,512,0.01921688848071628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,10240,0.06888266404469807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,1024,0.0031226666437255014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,12288,0.08471110794279312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,10240,0.016161777906947665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,7168,0.007657777931955125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,3584,0.01883288886811998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,64,0.005943111247486538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,512,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,768,0.029105779197480943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,4096,0.02799288762940301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,16384,0.04000799854596456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,3584,0.014440000057220459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,128,0.002191999927163124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,512,0.011024889018800525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,2048,0.006670222100284364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,768,0.01683466633160909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,4096,0.009756444229020013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,3072,0.007625777688291337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,65536,7.758740743001302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,65536,0.28848711649576825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,128,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,5120,0.01610577768749661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,5120,0.010703999963071613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,16384,0.020633776982625324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,5120,0.012735111018021902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,3072,0.018795554836591084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,65536,0.028443555037180584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,2048,0.007371555599901412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,2560,0.11796977784898545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,1536,0.010378667049937779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,1536,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,64,0.009303111169073317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,256,0.004908444566859139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,768,0.0028000000036425064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,16384,0.027588443623648748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,2048,0.00683822234471639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,768,0.003866666720973121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,512,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,768,0.005888888819350137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,3072,0.029208001163270738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,768,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,7168,0.04288622074657016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,2560,0.013743110828929476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,2560,0.026506667335828144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,3584,0.3087208800845676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,8192,0.014835556348164877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,512,0.006623999940024481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,2048,0.013949332965744866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,4096,0.009510222408506606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,5120,0.009323555562231276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,7168,0.032820443312327065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,10240,0.038667556312349104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,16384,0.0207306659883923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,3584,0.008683555656009251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,16384,0.01614488826857673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,64,0.004238222208287981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,12288,0.10833688577016194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,3584,0.03460266523891025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,16384,0.047451555728912354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,2560,0.006609777609507243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,768,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,256,0.014190221826235453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,64,0.005147555636035072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,2560,0.015821332732836407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,32,0.002514666567246119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,2048,0.005035555611054103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,5120,0.018908444378111098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,128,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,8192,0.021380444367726643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,6144,0.018013333280881245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,2048,0.006631999793979857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,768,0.01697688963678148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,12288,0.2269608974456787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,5120,0.008545777863926357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,3584,0.016598222984208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,3072,0.06896977954440646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,65536,0.06016355752944946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,2048,0.006302222195598815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,4096,0.01124533348613315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,10240,0.0273982220225864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,5120,0.10399999883439805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,10240,0.006608888920810487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,65536,0.22571288214789498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,7168,0.009697777529557547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,8192,0.02567111121283637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,128,0.004690666579537922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,16384,0.015576000014940897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,3072,0.02781244450145298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,256,0.005000000198682149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,32,0.0027466666781240036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,128,0.014447111222479077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,12288,0.06455022096633911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,64,0.001855111163523462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,8192,0.21395556131998697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,6144,0.02472355630662706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,7168,0.01648888819747501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,128,0.005962666538026597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,2048,0.012076444096035428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,2560,0.024492444263564214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,128,0.1637608872519599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,3584,0.01199288914601008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,12288,0.024455999334653217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,10240,0.023411555422676936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,6144,0.01349244432316886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,256,0.006612444503439798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,32,0.007352888584136963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,1536,0.00979911122057173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,5120,0.020671111014154222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,128,0.011358222199810876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,256,0.004925333377387789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,512,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,256,0.01277599980433782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,16384,0.020285333196322124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,4096,0.0051893335249688895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,10240,0.1445146666632758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,3584,0.029194666279686823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,7168,0.009698666632175446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,1536,0.006214222146405115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,5120,0.01479466590616438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,10240,0.016128000285890367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,256,0.004590222405062782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,1024,0.005506666584147348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,3584,0.09823999802271526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,256,0.005251555393139522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,2048,0.01276177747382058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,1536,0.017532444662517972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,3072,0.007272889216740926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,10240,0.01866933372285631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,2048,0.012554667062229581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,12288,0.031805333164003156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,768,0.006240889016124938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,2048,0.019931554794311523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,768,0.003212444484233856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,65536,0.13159999582502577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,512,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,2560,0.0069644442862934535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,4096,0.041730665498309664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,8192,0.04277599851290385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,3072,0.016163556112183463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,2048,0.006760889043410619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,65536,1.9955431620279949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,768,0.006314666734801398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,16384,0.39638755056593156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,3072,0.04271733429696825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,3072,0.01480977733929952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,6144,0.009017777939637503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,65536,0.05628089110056559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,128,0.014114666316244336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,6144,0.011335999601417117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,65536,0.07956710788938734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,1024,0.0031902222997612427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,2560,0.033333334657881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,32,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,5120,0.014059555199411182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,2560,0.016881777180565726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,512,0.005387555393907759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,2560,0.0287715560860104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,8192,0.033972445461485125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,3072,0.006928000185224745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,1536,0.015455111861228943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,65536,0.08793066607581244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,65536,0.3926604323916965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,8192,0.010998222563001843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,4096,0.1283066670099894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,7168,0.01221866657336553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,3072,0.0336808893415663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,512,0.2114000055525038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,64,0.0028159999185138275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,3584,0.0073626670572492815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,12288,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,10240,0.026476444469557867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,3072,0.025439111722840205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,16384,0.011911110745535957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,3584,0.012759110993809171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,1024,0.005951111101441913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,512,0.005305777821275923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,6144,0.014209777116775513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,2560,0.009693333672152625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,1536,0.021597334080272253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,8192,0.05023466547330221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,256,0.00490133340160052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,16384,0.020457777712080214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,32,0.005522666705979242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,10240,0.08585421906577216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,2048,0.015465777781274585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,3584,0.11430488692389594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,2048,0.026538666751649644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,5120,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,1024,0.018392889036072623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,3584,0.0087013335691558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,256,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,3584,0.06961244344711304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,2560,0.05781600210401747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,16384,0.028414222929212783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,2560,0.006978666616810693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,3072,0.00795999997191959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,128,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,5120,0.06421955426534016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,10240,0.011714666253990598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,1536,0.016166221764352586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,7168,0.015856888559129503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,3584,0.007622222105662028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,3584,0.009111111362775167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,65536,0.4657511181301541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,2560,0.008987555901209513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,6144,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,512,0.00565244464410676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,12288,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,10240,0.024308444725142583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,65536,0.03444800111982558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,5120,0.015965332587560017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,10240,0.09237777524524265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,4096,0.005883555445406172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,7168,0.015463110473420886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,6144,0.018731556004948087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,2048,0.006238222122192383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,128,0.014502222339312235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,6144,0.03693244523472256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,12288,0.013208000196350945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,2048,0.012582222620646158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,65536,0.8082515398661295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,5120,0.04474044508404202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,5120,0.025806221697065566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,65536,0.20148089196946886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,512,0.024534222152498033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,1536,0.040932443406846784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,65536,0.4664195378621419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,6144,0.17472178406185576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,256,0.004927999857399198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,5120,0.018576888574494254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,8192,0.0277502222193612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,2560,0.021112889051437378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,1024,0.007658667034573025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,7168,0.019592000378502738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,3072,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,3072,0.007837333612971837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,10240,0.028536889288160536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,5120,0.04334933227962918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,2560,0.022258667482270136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,1536,0.015795555379655626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,2560,0.013588444226317935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,5120,0.07678666379716662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,2048,0.1302328904469808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,4096,0.021976888179779053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,256,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,6144,0.01387555566098955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,1536,0.15652533372243246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,12288,0.011570666399266986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,12288,0.013070222404268054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,128,0.0052328887912962176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,16384,0.06421955426534016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,128,0.0042053332759274375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,3072,0.0276826669772466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,512,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,3072,0.08283555507659912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,10240,0.043495999442206494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,6144,0.005947555518812603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,4096,0.012674666941165924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,512,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,2048,0.015968888998031616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,128,0.002199999988079071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,7168,0.01684977776474423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,10240,0.0875848862859938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,12288,0.009321778184837764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,8192,0.017642666896184284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,64,0.002165333264403873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,5120,0.06395644611782499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,6144,0.3826666673024495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,3584,0.011066666907734342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,10240,0.203475554784139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,8192,0.03532888823085361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,3584,0.012069332930776807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,65536,0.08201155397627088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,1536,0.005944889038801193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,3584,0.01585422290696038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,5120,0.025076443950335186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,7168,0.03830844495031569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,7168,0.01826844447188907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,8192,0.011075555450386472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,3072,0.015565334094895257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,2048,0.01608533329433865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,256,0.015125332607163323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,8192,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,6144,0.01350933313369751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,8192,0.08506400055355495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,3072,0.011669333610269757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,768,0.006565333240562015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,10240,0.012506666282812754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,12288,0.026996443669001263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,6144,0.07502755853864881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,2048,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,3072,0.007257777783605788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,4096,0.04072355561786228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,64,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,768,0.05416799916161431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,5120,0.020966221888860066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,10240,0.004861333303981357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,32,0.002833777831660377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,512,0.004879999905824661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,768,0.007346666521496243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,8192,0.01776355504989624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,256,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,1536,0.031290666924582586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,3584,0.008028444316652086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,12288,0.03846577803293864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,4096,0.015573332707087198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,128,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,256,0.004919111314747068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,768,0.018869333797030978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,65536,0.5874106619093153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,10240,0.016197333733240765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,7168,0.0839315586619907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,2560,0.09928088717990452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,7168,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,4096,0.017204445269372728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,8192,0.0354639987150828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,1536,0.016514667206340365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,65536,0.009482666850090027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,2048,0.06677955389022827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,16384,0.22938844892713758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,3072,0.007751999629868402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,768,0.005937777873542573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,12288,0.07537511322233412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,3072,0.021709332863489788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,3584,0.007637333538797166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,10240,0.02199999988079071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,3072,0.11045510901345147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,2560,0.0074035558435651995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,2048,0.006931555353932911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,128,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,2048,0.011730666789743634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,2560,0.010031999813185798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,7168,0.01553066737122006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,10240,0.014748444159825643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,5120,0.008303999900817871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,12288,0.013768889009952545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,10240,0.011700444751315646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,8192,0.03402577837308248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,4096,0.02608799934387207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,64,0.0018595556418100994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,3584,0.007978666987684038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,4096,0.01308444473478529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,12288,0.020402666595247056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,64,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,10240,0.016591999265882704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,3584,0.019324445062213473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,1024,0.013009777499569787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,1024,0.003181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,256,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,6144,0.0258933338854048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,16384,0.017284444636768766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,3584,0.010004444254769219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,6144,0.026914666096369427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,4096,0.00793688909875022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,5120,0.0083155557513237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,10240,0.016921778519948322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,6144,0.10257866647508408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,3584,0.019524445136388142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,16384,0.06164444155163235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,64,0.0019031111150979996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,8192,0.007661333514584436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,3584,0.09023822016186184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,2048,0.004583999928500918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,4096,0.006362666686375936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,128,0.005296000176005893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,32,0.009672889278994666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,6144,0.024852444728215534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,32,0.001869333287080129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,65536,0.00980000032318963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,12288,0.014089778065681458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,32,0.009683555199040307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,32,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,5120,0.010987555815113915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,768,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,1024,0.01722577710946401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,65536,0.09228889147440593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,128,0.007998222278224098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,1024,0.021918222308158875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,1024,0.009215999808576372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,5120,1.5257333119710286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,6144,0.016157334049542744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,6144,0.012823111481136747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,12288,0.025719111164410908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,4096,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,7168,0.00979288915793101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,1536,0.006317333214812809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,32,0.017265778448846605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,1536,0.008527999950779809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,512,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,3584,0.021867555048730638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,256,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,5120,0.012856889102194043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,12288,0.021801778011851843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,2560,0.026199110680156287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,12288,0.01924711134698656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,4096,0.007956444389290279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,12288,0.019918221566412184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,16384,0.12068088849385579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,3584,0.007481777833567725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,2048,0.019540444016456604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,16384,0.017091555727852713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,16384,0.03763377666473389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,32,0.0021297778520319196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,3584,0.04157155421045091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,256,0.007654222349325816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,768,0.010033778018421596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,8192,0.012411555482281579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,3072,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,768,0.02129688858985901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,12288,0.24914044804043242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,8192,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,128,0.00216977774269051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,12288,0.30163555675082737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,512,0.011391999820868174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,256,0.022758222288555566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,3072,0.014123555686738757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,7168,0.010041777458455827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,16384,0.014687110980351767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,256,0.016203555795881484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,768,0.008638222184446123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,5120,0.02624977793958452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,768,0.003847111016511917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,65536,0.07843289110395643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,128,0.011351111034552256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,768,0.015494222442309061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,2048,0.03239377670817905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,7168,0.010335110955768162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,4096,0.011018666956159802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,2560,0.017529777354664274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,12288,0.031629333893458046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,10240,0.049032890134387545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,10240,0.6090755462646484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,6144,0.023355555203225877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,3072,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,12288,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,64,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,512,0.005052444421582752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,512,0.0063475556671619415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,8192,0.057532442940606006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,64,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,128,0.014479110638300577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,65536,0.023352000448438857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,1024,0.00628977765639623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,10240,0.013456000222100152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,1536,0.02197066611713833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,5120,0.04392533169852363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,8192,0.005873777800136142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,768,0.01921955578856998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,5120,0.013357333011097379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,7168,0.03905155592494541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,1024,0.006242666807439592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,1536,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,3584,0.01477777792347802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,8192,0.06160800324545967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,5120,0.2348897722032335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,7168,0.011057777537239922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,3584,0.02197955548763275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,16384,0.047925333182017006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,6144,0.007589333587222629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,4096,0.008396444221337637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,2048,0.004218666503826777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,7168,0.016170667277442086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,8192,0.01107911103301578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,3584,0.01460444430510203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,16384,0.0396053327454461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,1536,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,128,0.014156444205178155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,2560,0.013191111385822296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,5120,0.011728000309732227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,128,0.014446222119861178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,3072,0.027436445156733196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,256,0.007296000089910295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,8192,0.018258665998776753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,3584,0.032913777563307024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,16384,0.00869422240389718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,16384,0.05474755499098036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,768,0.04134755664401584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,2560,0.006600889066855113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,8192,0.05124800072775947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,12288,0.021329777108298406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,32,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,2560,0.01312800000111262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,7168,0.06651733318964641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,65536,0.18795200188954672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,8192,0.010593777729405297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,6144,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,768,0.017190222938855488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,512,0.0052151112920708126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,3584,0.014820444915029736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,10240,0.20141422748565674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,2560,0.004998222407367495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,6144,0.040279110272725425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,10240,0.004870222260554631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,3072,0.07654133107927111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,8192,0.06456177764468722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,1536,0.011391110718250275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,16384,0.04569777846336365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,7168,0.04781066709094577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,2560,0.021427555216683283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,1536,0.027414222558339436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,1536,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,2048,0.012103111379676394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,256,0.005192888693677055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,1536,0.09020266930262248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,4096,0.026368889543745253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,65536,0.0973022249009874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,16384,0.008293333152929941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,3584,0.038482666015625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,1536,0.08381333616044785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,256,0.016807110773192514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,64,0.007582222421964009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,2048,0.006966222491529252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,4096,0.010043555663691627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,3072,0.004907555464241239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,1024,0.005990222096443176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,12288,0.026706667410002813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,5120,0.0314631097846561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,512,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,8192,0.023197333017985027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,256,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,65536,0.08487821949852838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,4096,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,5120,0.008414222134484185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,5120,0.016509332590632968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,6144,0.06842488712734647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,10240,0.05446755554940966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,1024,0.014114666316244336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,3072,0.012476444244384766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,128,0.0022293333378103045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,10240,0.9632426367865669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,6144,0.18385777208540174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,2560,0.0222746663623386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,1024,0.02165599995189243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,8192,0.1785137785805596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,32,0.005205333232879639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,32,0.0019262221952279408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,2048,0.004320888883537716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,8192,0.010835555692513784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,256,0.016864889197879367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,6144,0.012240889171759287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,65536,0.067885332637363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,3072,0.007318221860461765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,3072,0.014335999886194864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,7168,0.0083155557513237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,1536,0.007104000283612146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,12288,0.12936177518632677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,2048,0.0042204442951414315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,10240,0.024170666933059692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,512,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,1536,0.014117333624098035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,65536,0.2887937757703993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,4096,0.01925511161486308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,3072,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,3584,0.00813244448767768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,2048,0.006634666687912411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,64,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,256,0.021898667017618816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,2560,0.007224000162548489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,3584,0.00693866651919153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,256,0.01513066722287072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,16384,0.029321776496039495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,768,0.003499555504984326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,768,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,12288,0.178911116388109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,16384,0.011257777611414591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,768,0.01717688971095615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,2560,0.006919999917348226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,32,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,1536,0.016530666086408827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,8192,0.017161778277821012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,256,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,12288,0.02122666603989071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,10240,0.02074044446150462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,16384,0.025055110454559326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,1024,0.006263999889294307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,16384,0.015299555328157214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,3072,0.02713866697417365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,10240,0.010324444207880232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,128,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,10240,0.052552888790766396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,2048,0.006639110959238476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,1024,0.015234667393896313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,10240,0.018623110320832994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,32,0.005036444299750858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,2560,0.007761778102980719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,2048,0.006607111129495833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,16384,0.07396444347169664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,512,0.014759999182489185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,65536,0.12981332672966853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,16384,0.016143111719025504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,7168,0.386757320827908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,2048,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,12288,0.02811288833618164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,6144,0.008300444318188561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,128,0.014139556222491793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,2048,0.006614222294754452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,7168,0.030156443516413372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,512,0.02693422304259406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,2560,0.015454222758611044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,5120,0.05746044715245565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,32,0.00562755556570159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,3072,0.021356445219781663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,2048,0.02087822225358751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,8192,0.015164444843928019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,768,0.0038293335172865125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,1024,0.011795555551846823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,32,0.008974221845467886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,2048,0.03013066781891717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,6144,0.9190408918592664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,8192,0.251888010236952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,4096,0.007976888782448238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,512,0.015032000011867948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,10240,0.035622222556008235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,512,0.016395555602179635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,128,0.010320888625250923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,4096,0.017864000466134813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,256,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,2048,0.014105778601434497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,2048,0.006592000110281839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,1024,0.00757422215408749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,4096,0.007320000065697565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,4096,0.011012444065676795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,512,0.01733955575360192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,5120,0.008539555801285638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,2560,0.006616888774765863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,3072,0.04900177650981479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,3072,0.00777688870827357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,6144,0.02314666741424137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,65536,0.1027217772271898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,128,0.006272000157170826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,128,0.004577777865860197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,4096,0.008648888932334052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,512,0.004991111242108875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,5120,0.04690666662322151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,6144,0.01609155535697937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,6144,0.013256889250543384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,12288,0.09188977877298991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,8192,0.03690133492151896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,16384,0.04025333457522922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,4096,0.008268444074524773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,10240,0.03539289037386576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,6144,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,128,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,8192,0.012073777616024017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,7168,0.012058667010731168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,4096,0.00979555563794242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,768,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,65536,0.06145955456627739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,256,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,128,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,32,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,65536,0.08672266536288792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,256,0.0214008887608846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,12288,0.027601778507232666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,256,0.012156444291273752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,6144,0.04122399952676561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,7168,0.02649333410792881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,6144,0.044657776753107704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,6144,0.015467555986510383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,128,0.028247998820410833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,1536,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,10240,0.01274577776590983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,512,0.008640888664457533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,512,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,8192,0.013479111095269522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,3072,0.006911110960774952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,12288,0.14249244001176622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,5120,0.1439173354042901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,3584,0.007207999626795451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,768,0.004913777940803104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,2560,0.04485333296987745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,65536,0.4676648775736491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,256,0.0024951110697454875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,768,0.017044444878896076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,65536,0.178984006245931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,4096,0.026185777452256944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,256,0.004912000149488449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,1536,0.005688000056478713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,6144,0.024799111816618178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,3072,0.05799733267890083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,12288,0.013639111485746173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,12288,0.2171342240439521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,6144,0.5933973524305556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,64,0.0025439999169773525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,4096,0.06880533032947116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,7168,0.01205599970287747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,1536,0.012913777596420713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,3072,0.007670222057236566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,256,0.02437866727511088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,8192,0.04042844308747186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,512,0.015040000279744467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,32,0.002151111140847206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,65536,0.208186666170756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,768,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,7168,0.004534222185611725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,128,0.0024871110088295406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,256,0.003504000190231535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,7168,0.03645688957638211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,8192,0.0350488887892829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,12288,0.03386133246951633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,16384,0.02476177778508928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,256,0.027331555883089702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,1024,0.029885331789652508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,6144,0.012727999852763282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,512,0.012794666820102267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,4096,0.01204355557759603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,256,0.014469333820872836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,5120,0.013044444223244986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,3072,0.010327999790509542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,2560,0.025221332907676697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,256,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,128,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,32,0.005689777847793367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,65536,0.6143609152899848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,512,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,3584,0.011352889239788055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,5120,0.009325332939624786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,2048,0.006577777779764599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,3584,0.013635555903116861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,5120,0.029342220889197454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,8192,0.022520000735918682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,8192,0.02360444433159298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,2048,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,2048,0.009895111123720804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,8192,0.038506666819254555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,128,0.00282222218811512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,8192,0.05685777796639336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,12288,0.02648799949222141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,12288,0.013659555878904132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,1024,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,2560,0.007618666523032718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,16384,0.02823822365866767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,1024,0.004886222382386525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,5120,0.00830577810605367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,2048,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,5120,0.012471111284361945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,1024,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,256,0.016528889536857605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,65536,0.17371911472744414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,1536,0.004550222307443619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,2048,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,12288,0.01926666663752662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,12288,0.04804977774620056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,10240,0.02015466656949785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,512,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,256,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,3072,0.022138666775491502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,512,0.008788444101810455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,4096,0.025015999873479206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,256,0.005214222189452913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,65536,0.06614844666586982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,768,0.010465777582592435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,1024,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,65536,0.08277866575453016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,16384,0.03489511211713155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,1024,0.003170666595300039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,1536,0.0034977777136696708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,768,0.006275555739800136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,128,0.005530666559934616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,5120,0.023648000425762598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,3584,0.01516355574131012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,1024,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,10240,0.045462220907211304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,256,0.00979555563794242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,16384,0.25189420912000865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,7168,0.018198221921920776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,512,0.0038720000949170855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,12288,0.014779556128713818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,8192,0.03517599900563558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,2560,0.006625777731339137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,1536,0.0038737778862317405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,1536,0.0034968890249729156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,128,0.00327822214199437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,4096,0.06749244530995686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,6144,0.01943644384543101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,5120,0.008623110751310984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,32,0.0021395554973019496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,128,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,1024,0.006251555350091722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,8192,0.013917333549923368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,8192,0.01650399963061015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,768,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,512,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,1536,0.0035137778355015647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,10240,0.14391467306349012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,4096,0.03047377864519755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,64,0.005167110926575131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,2560,0.008007110820876228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,16384,0.0270506673389011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,64,0.005686222265164058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,2560,0.007387555307812161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,512,0.0038408889538711975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,6144,0.06872089041603936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,512,0.006959111326270633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,5120,0.01587288909488254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,1536,0.0038373333712418876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,768,0.014153778553009033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,256,0.019911999503771465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,256,0.057348443402184375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,64,0.00553688903649648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,10240,0.08474933438830906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,32,0.006063111126422882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,768,0.007634667058785756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,256,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,2048,0.02679555614789327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,3072,0.00795288880666097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,5120,0.01574577722284529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,12288,0.02747111188040839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,64,0.0018008889423476325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,16384,0.10778933101230198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,256,0.006223111102978389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,6144,0.015684444043371413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,768,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,12288,0.0602746672100491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,3584,0.004839111119508743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,4096,0.007637333538797166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,256,0.011232888533009423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,64,0.0019057778020699818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,16384,0.1723031070497301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,1536,0.03444800111982558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,65536,0.20480177137586805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,2048,0.028913779391182795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,32,0.00216977774269051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,4096,0.25843199094136554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,6144,0.019902222686343722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,128,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,7168,0.037329776419533625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,6144,0.011062222222487131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,3584,0.012712889247470431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,32,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,4096,0.11698488394419353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,7168,0.016999999682108562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,256,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,3072,0.031509333186679415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,2048,0.013762666947311826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,6144,0.010115555591053432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,8192,0.02109955582353804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,64,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,6144,0.016542222764756944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,12288,0.013770666387346057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,256,0.002149333349532551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,64,0.0028719999310043124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,4096,0.028452446063359577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,8192,0.03244444396760728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,12288,0.02667288813326094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,3584,0.03583911061286926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,16384,0.005389333185222413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,8192,0.03906044363975525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,65536,0.008616000413894653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,3584,0.007530666887760162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,768,0.02590399980545044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,3584,0.007628444168302748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,2560,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,8192,0.11479111512502034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,2048,0.04519911275969612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,256,0.004877333425813251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,7168,0.01238844460911221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,256,0.00564177789621883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,5120,0.03226666649182638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,7168,0.022397332721286353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,1024,0.021665778425004747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,512,0.03461688756942749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,12288,0.023929778072569106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,8192,0.014175999495718213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,2048,0.012794666820102267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,5120,0.07791911231146918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,12288,0.03535822365019057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,512,0.007632888853549957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,5120,0.03798310955365499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,65536,0.668283568488227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,16384,0.023676445086797077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,512,0.0041724443435668945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,1536,0.022355554832352534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,3072,0.014599111345079211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,10240,0.026399999856948853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,768,0.003939555750952827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,10240,0.08421244223912557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,128,0.005504000104135937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,32,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,65536,0.10264977481630112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,2048,0.012888888518015543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,512,0.005370666583379109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,32,0.007672000262472365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,2048,0.00599644457300504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,12288,0.05135200089878506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,5120,0.013982221484184265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,5120,0.0339902208911048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,12288,0.025752888785468206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,7168,0.011914666328165265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,6144,0.00906133320596483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,2560,0.059451553556654185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,2560,0.00666844430896971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,512,0.01776622235774994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,64,0.005290666802061929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,16384,0.020616888999938965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,3072,0.03370577759212918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,16384,0.03247022297647264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,4096,0.008310221963458592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,8192,0.01108266661564509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,64,0.005935110979610019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,12288,0.08066222402784559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,256,0.0090515555606948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,512,0.02885599931081136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,256,0.0028106667515304354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,5120,0.04986844460169474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,8192,0.014735110931926303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,64,0.006550222221348021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,16384,0.010023111270533668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,768,0.016680889659457736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,3584,0.007302222152551015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,10240,0.04115555683771769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,3072,0.017196445001496207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,1536,0.014110222458839417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,512,0.006592000110281839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,256,0.014124444789356656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,16384,0.08859733078214858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,16384,0.023991111252042983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,3584,0.011414222419261932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,4096,0.005000888887378904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,1024,0.009316444396972656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,32,0.0018666666001081467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,10240,0.015439111325475903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,64,0.0054551110499435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,3072,0.06036444505055746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,2048,0.017218665944205392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,7168,0.011363555159833698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,8192,0.01347644461525811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,8192,0.016523554921150208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,7168,0.015318221516079374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,10240,0.02569955587387085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,7168,0.016216889023780823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,512,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,3584,0.01481599940194024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,128,0.009762666291660732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,2560,0.013763555222087435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,3584,0.016179554992251925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,3072,0.01204355557759603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,512,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,8192,0.03222489025857713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,1024,0.006993778049945831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,1024,0.0028293333533737394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,3072,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,5120,0.008649778034951951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,512,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,768,0.02881511052449544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,1024,0.003899555653333664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,7168,0.02421244482199351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,32,0.005622222191757626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,128,0.006956444846259222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,512,0.0053075556125905775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,256,0.005281777845488654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,768,0.01628888812330034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,16384,0.04463111029730903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,5120,0.011720000041855706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,128,0.013791110780504016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,3584,0.007536888950400882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,768,0.0076044441925154785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,65536,0.07204711437225342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,64,0.005538666827811136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,3072,0.015178667174445258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,1536,0.018579555882347953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,10240,0.015050666199790107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,65536,0.21044533782535127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,1536,0.003522666792074839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,64,0.002206222257680363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,10240,0.12558400630950928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,1536,0.006222222414281633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,3584,0.03207555413246155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,512,0.00509066672788726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,12288,0.026399999856948853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,10240,0.17608443895975748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,768,0.017235555582576327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,7168,0.007320889168315464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,65536,0.1874391105439928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,32,0.021976888179779053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,8192,0.37705689006381565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,256,0.0023608888602919048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,12288,0.009703999592198266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,2560,0.02624977793958452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,32,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,1536,0.006767999794748094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,3584,0.0076657773719893555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,5120,0.027404444085227117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,8192,0.5991831355624729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,128,0.033505777517954506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,1536,0.011114666859308878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,2048,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,16384,0.008738666772842407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,2560,0.008968888885445064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,10240,0.019241778386963736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,5120,0.17950667275322807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,4096,0.023022222850057814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,256,0.0026008888251251644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,12288,0.07610044214460585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,7168,0.07840089003245036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,2048,0.019552888141738046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,7168,0.023961777488390606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,2560,0.008979555633332994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,6144,0.006919999917348226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,2560,0.05928444200091892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,12288,0.02001422146956126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,8192,0.014027555783589682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,12288,0.2117582162221273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,1536,0.006211555666393704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,65536,0.15442932976616755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,3072,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,32,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,64,0.012044444680213928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,2048,0.018632888793945312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,5120,0.004594666676388847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,65536,1.8999875386555989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,128,0.006618666566080517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,1024,0.007431999676757389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,3072,0.01715555621518029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,64,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,7168,0.013788444300492605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,16384,0.07467110951741536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,128,0.008016000191370646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,1536,0.01717155509524875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,1536,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,2048,0.02300888962215847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,7168,0.14578133159213597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,2560,0.010059555371602377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,10240,0.05274577935536703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,3584,0.11901422341664632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,4096,0.02440711193614536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,6144,0.009299555586444007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,65536,0.11319022708468968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,12288,0.2401111125946045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,32,0.0021173333128293357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,1024,0.005596444424655702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,256,0.016487111647923786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,2560,0.006809777683681912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,768,0.007260444263617198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,2560,0.008965333302815756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,3072,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,65536,0.08244533008999295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,3584,0.01922044489118788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,3072,0.01651733285850949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,4096,0.02102577851878272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,5120,0.03535644544495477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,128,0.0028577778074476454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,10240,0.018023111753993563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,2560,0.009704888694816166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,32,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,1536,0.006587555425034628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,12288,0.019487111104859244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,8192,0.01184177812602785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,5120,0.008545777863926357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,3584,0.014704888065656027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,3584,0.03781866696145799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,65536,0.09488977988560994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,128,0.013774221969975365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,32,0.006232000059551663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,2048,0.009912889036867354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,4096,0.007886221839321984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,256,0.005761777775155173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,32,0.0021262222694026097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,8192,0.07659644550747342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,2048,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,65536,0.07917155822118123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,32,0.002021333409680261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,128,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,5120,0.00867555538813273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,768,0.02936799989806281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,5120,0.011700444751315646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,1536,0.027926223145590887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,6144,0.06853510936101277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,1024,0.009325332939624786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,7168,0.02238666680124071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,64,0.0018604445374674266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,32,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,2048,0.005787555542257097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,32,0.009663999908500248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,512,0.035882665051354304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,6144,0.01835466590192583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,768,0.011055111057228513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,3072,0.26916445626152885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,12288,0.05797244442833794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,4096,0.007610666255156199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,4096,0.009912889036867354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,5120,0.008989333278603023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,8192,0.01553600033124288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,768,0.006582222051090664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,128,0.002163555473089218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,2048,0.043952001465691455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,768,0.0069066666894488865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,32,0.005290666802061929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,1536,0.02978488802909851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,4096,0.01517599986659156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,2560,0.012847111456924014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,2560,0.004520888957712385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,3584,0.007622222105662028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,768,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,12288,0.018963555494944256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,12288,0.024088889360427856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,256,0.007963555554548899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,1024,0.014783110883500842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,12288,0.033702221181657575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,128,0.00796622203456031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,16384,0.8886826833089193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,7168,0.03320533368322585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,1024,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,128,0.00905688852071762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,1024,0.005929777605666055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,256,0.00527555536892679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,7168,0.007719111111429002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,2560,0.5697253545125326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,768,0.006279110908508301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,256,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,2048,0.10239999824100071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,3072,0.016573333077960543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,2048,0.012831110921170978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,10240,0.09172177977032131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,6144,0.017109332813156974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,2048,0.00628977765639623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,10240,0.8322790993584527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,1024,0.004474666797452503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,128,0.00793599999613232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,5120,0.02401777770784166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,768,0.005912000106440649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,12288,0.03583555420239767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,6144,0.02204977803760105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,1024,0.027497778336207073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,5120,0.15462222364213732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,2048,0.0086666668454806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,6144,0.009263111485375298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,3584,0.026205332742797002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,512,0.002534222271707323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,7168,0.2780826621585422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,1024,0.0107351111041175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,1024,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,7168,0.037793778710895114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,6144,0.013811555173661975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,16384,0.08949244684643215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,1536,0.005355555564165115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,7168,0.010658666491508484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,1536,0.006129777679840724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,16384,0.03063911199569702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,3072,0.006973333656787872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,65536,0.18776444594065347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,1536,0.006456888798210356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,1536,0.021701332595613267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,512,0.02660177813635932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,768,0.0028133332315418455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,5120,0.025067556235525344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,5120,0.028102222416136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,8192,0.010446222292052375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,1536,0.006144000010357962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,1536,0.01889866590499878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,768,0.029439111550649006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,4096,0.006568888823191325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,32,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,65536,0.04455822375085619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,2560,0.034895999564064875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,3072,0.014114666316244336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,64,0.006207110981146495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,16384,0.05083022183842129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,7168,0.009691555466916826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,128,0.00591111100382275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,1536,0.0038533334930737815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,7168,0.012747555143303342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,32,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,12288,0.13011822435590956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,32,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,12288,0.06249688731299507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,32,0.007668444679843054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,128,0.036249776681264244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,2560,0.006647999915811751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,2560,0.006709333509206772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,4096,0.015136889285511441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,128,0.003526222374704149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,7168,0.1611351172129313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,768,0.019825778073734708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,256,0.022023111581802368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,512,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,3584,0.1646746661927965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,65536,0.10368000136481391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,65536,0.3041448858049181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,768,0.039706667264302574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,2560,0.020118221640586853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,128,0.029166221618652344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,10240,0.01958666741847992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,16384,0.019433778193261888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,1536,0.009319999979601966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,12288,0.014837332897716098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,8192,0.04714222086800469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,10240,0.012064889073371887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,12288,0.03586133321126302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,65536,0.1564026673634847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,8192,0.016548444827397663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,1024,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,128,0.013447999954223633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,2048,0.012768000364303589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,3072,0.007423111134105259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,128,0.006211555666393704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,128,0.008447111480765873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,2048,0.024117334021462336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,512,0.01254666679435306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,65536,0.11569955613878037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,3584,0.008085333638721043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,8192,0.0144195556640625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,4096,0.008570666942331526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,12288,0.01648177703221639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,512,0.013456000222100152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,12288,0.05884444713592529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,8192,0.010051555931568146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,1024,0.0035235554807715942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,128,0.004607111215591431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,5120,0.01719999975628323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,256,0.004918222212129169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,768,0.012291555603345236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,256,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,256,0.006072888771692912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,16384,0.01651822196112739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,65536,0.044159111049440175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,1536,0.015464888678656684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,32,0.016252444850073922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,32,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,8192,0.01716088917520311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,16384,0.028775112496482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,512,0.011372444530328115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,65536,0.1722400055991279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,6144,0.03323288758595785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,12288,0.44377512402004665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,2560,0.010030222435792288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,1536,0.006060444646411472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,6144,0.04323022233115303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,768,0.013780444032616086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,3584,0.016528000434239704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,256,0.019259555472267997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,5120,0.004267555558019214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,1536,0.008283555507659912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,65536,0.07640355825424194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,128,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,8192,0.01108266661564509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,2048,0.009704888694816166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,10240,0.012358222570684222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,128,0.012335999972290464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,2560,0.015135111080275642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,8192,0.014146667387750415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,65536,0.5967973073323568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,768,0.018587556150224473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,768,0.008636444807052612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,3584,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,8192,0.02880177895228068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,65536,0.08674844768312241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,1024,0.0031902222997612427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,256,0.0021537778278191886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,128,0.002151111140847206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,512,0.013785777820481194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,5120,0.0063182223174307085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,65536,0.07692799965540568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,4096,0.00793333351612091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,7168,0.05013333426581489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,4096,0.007298666569921706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,2560,0.009197333620654212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,6144,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,16384,0.05080088973045349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,65536,0.11914666493733723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,768,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,2048,0.015795555379655626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,7168,0.009663110805882348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,3072,0.006969778074158563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,7168,0.03993777765168084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,1536,0.004182222402758068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,4096,0.01181244436237547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,65536,0.3277013301849365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,65536,0.09182311428917779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,16384,0.03671200076738993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,16384,0.5019084612528483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,2560,0.009362666971153682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,7168,0.09926311175028484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,5120,0.008960000342792934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,10240,0.03246488836076524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,12288,0.01757688820362091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,6144,0.04553688897026909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,8192,0.010374222364690568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,1536,0.15614488389756945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,65536,0.3223493364122179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,1536,0.03751111030578613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,16384,0.024513777759340074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,2560,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,2048,0.023659555448426142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,2048,0.00425866660144594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,5120,0.11241155200534397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,3584,0.029747555653254192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,768,0.013840888937314352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,32,0.001998222122589747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,16384,0.017128888103697035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,7168,0.06958578030268352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,7168,0.009838222629494136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,5120,0.004949333353175057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,64,0.010761777559916178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,5120,0.010040889183680216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,3584,0.004604444321658877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,4096,0.02913244565327962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,4096,0.012154666913880242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,7168,0.04996088809437222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,4096,0.0655226641231113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,3584,0.015799999237060547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,65536,0.015642666154437594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,1024,0.007895111209816402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,16384,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,128,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,512,0.01520088811715444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,4096,0.017443555924627516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,65536,0.44550667868720156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,4096,0.011758222348160215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,32,0.0028293333533737394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,65536,0.08247911267810397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,64,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,1024,0.016906667086813185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,6144,0.01421955558988783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,256,0.015881778465376962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,16384,0.028872887293497723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,2048,0.00998399986161126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,512,0.003220444545149803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,512,0.007965332931942409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,128,0.004932444542646408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,2048,0.014765333798196582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,2048,0.006616888774765863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,1536,0.011428444749779172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,2048,0.021272889441914026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,64,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,768,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,1024,0.006014222072230444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,512,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,256,0.003154666680428717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,128,0.01036711119943195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,256,0.10343377457724677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,128,0.004911111046870549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,1536,0.025024889243973628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,1536,0.006245333287451003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,7168,0.05042666527960035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,16384,0.02851288848453098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,65536,0.523031128777398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,12288,0.2675066788991292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,65536,1.7384719848632812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,2048,0.004931555440028508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,2048,0.01718488832314809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,8192,0.042640888028674655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,7168,0.02459200057718489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,768,0.01272266689274046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,2560,0.04103644357787238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,8192,0.08763200044631958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,12288,0.026405332816971674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,3072,0.017853332890404593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,2560,0.024002666274706524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,6144,0.015219555960761176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,4096,0.013368888861603208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,4096,0.007931555310885111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,512,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,128,0.002187555655837059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,1024,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,2048,0.0041848888827694785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,6144,0.0503786669837104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,16384,0.08834755420684814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,7168,0.017625777257813346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,1024,0.019920888874265883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,3072,0.030481778913074072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,12288,0.03479377759827508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,256,0.002716444432735443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,4096,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,2048,0.07035377952787611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,3072,0.01051288925939136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,1536,0.008619555168681675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,6144,0.031164444155163232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,7168,0.01607288916905721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,128,0.0021911110315057966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,128,0.005569777968857024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,65536,0.14605511559380427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,16384,0.041835556427637734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,8192,0.030528889762030706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,2048,0.010671999719407825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,10240,0.012216889195972018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,10240,0.08682399988174438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,512,0.005946666830115848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,10240,0.019533332851197984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,6144,0.11986666255527073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,768,0.0040835556056764396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,5120,0.11258933279249404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,3584,0.009893333746327294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,6144,0.007053333024183909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,16384,0.28671643469068736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,2048,0.018603555030292936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,12288,0.03392088744375441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,2048,0.00692622239391009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,768,0.007279111279381647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,4096,0.026749332745869953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,16384,0.028802666399214003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,6144,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,12288,0.018601778480741713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,256,0.016207999653286405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,2048,0.1136595540576511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,1024,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,768,0.0031822222388452957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,768,0.009280888570679558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,8192,0.015541333291265698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,256,0.00184888889392217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,1536,0.017058667209413316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,3584,0.11855022112528484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,7168,0.015051555302408008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,10240,0.00554666668176651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,3584,0.02389777700106303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,10240,0.03216977914174398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,128,0.01493866741657257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,10240,0.011731555892361535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,2048,0.017909333109855652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,3584,0.013418667018413544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,128,0.0021537778278191886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,6144,0.014774221513006421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,3072,0.036102222071753606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,2560,0.008884444832801819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,7168,0.037618666887283325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,8192,0.11586044894324408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,1024,0.042832887834972806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,128,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,1536,0.006203555398517185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,8192,0.011779555843936073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,7168,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,256,0.012383999923865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,64,0.001871111078394784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,1024,0.006981333096822103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,8192,0.019276445110638935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,6144,0.026133333643277485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,3072,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,1536,0.013809777796268463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,4096,0.010564444793595208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,12288,0.02612977723280589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,10240,0.005538666827811136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,6144,0.012087110843923358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,8192,0.013786666923099093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,8192,0.030272887812720403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,64,0.00519822206762102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,8192,0.06530933247672187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,3072,0.026404443714353774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,2560,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,7168,0.015816888875431485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,512,0.01313688854376475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,5120,0.016521778371598985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,256,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,65536,0.05541244480344984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,3072,0.011009777585665384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,10240,0.04113066527578566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,3584,0.007648888561460707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,2048,0.006457777900828256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,5120,0.0715342230266995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,5120,0.01754044493039449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,768,0.01889155473974016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,6144,0.008628444539176093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,16384,0.046569777859581836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,12288,0.03223022156291538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,256,0.009373333719041612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,768,0.019274666905403137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,768,0.014491554763582019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,768,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,5120,0.014646222194035849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,2560,0.012463111016485425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,12288,0.04083377785152859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,16384,0.04857866631613838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,8192,0.011600889265537262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,6144,0.01923822197649214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,1024,0.014485332700941296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,65536,0.12188443872663711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,5120,0.013754666679435305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,2560,0.2316497696770562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,65536,0.06351288821962145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,4096,0.007727999654081132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,10240,0.015806222955385845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,65536,0.15306133694118923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,2048,0.006292444550328785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,10240,0.012107555237081317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,2048,0.014717333846622042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,10240,0.5152391327752007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,7168,0.006262222097979651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,4096,0.04071999920739068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,1024,0.019560888409614563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,4096,0.04313866628540886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,8192,0.021648888786633808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,7168,0.011050666371981302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,65536,0.07022222545411852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,1536,0.01479022204875946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,5120,0.017830222845077515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,3072,0.009912889036867354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,768,0.005715555614895291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,1024,0.0037644443412621817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,2048,0.0041999999019834726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,128,0.004592888885074192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,256,0.010394666757848527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,10240,0.03093866507212321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,7168,0.00796977761718962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,4096,0.10298399792777167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,16384,0.12699111302693686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,1536,0.0037235555549462638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,5120,0.010016000105275048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,4096,0.010264889233642155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,5120,0.01204177737236023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,3584,0.027649778458807204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,2560,0.010547555155224271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,128,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,10240,0.03084889054298401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,1024,0.005986666513813867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,5120,0.07379466957516141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,256,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,7168,0.2198399967617459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,8192,0.042141334878073804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,6144,0.021711111068725586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,5120,0.015832000308566622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,2560,0.03657955593532986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,10240,0.028959999481836956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,64,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,6144,0.024805333879258897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,7168,0.009322666459613377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,2048,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,1536,0.011729777687125735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,1536,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,16384,0.016122667325867545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,2048,0.01757688820362091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,2048,0.006663110935025745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,128,0.01033866653839747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,4096,0.012118221984969245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,12288,0.5013813442654079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,128,0.028112000889248315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,3584,0.012107555237081317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,4096,0.011388444238238864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,5120,0.020957332518365648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,10240,0.07805244127909343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,1536,0.009905777871608734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,768,0.017893332574102614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,64,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,8192,0.011374221907721626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,3072,0.007630222373538547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,3584,0.026425777210129633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,65536,0.39385244581434464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,8192,0.004253333227501975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,3584,0.010403555300500657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,128,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,6144,0.01717333330048455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,1536,0.06760444243748982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,256,0.014482667048772177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,10240,0.011109333071443768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,16384,0.005209777918126848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,2560,0.007328000333574083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,2048,0.007009777757856581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,64,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,32,0.004218666503826777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,65536,0.889900419447157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,2048,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,8192,0.018184888694021437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,8192,0.01107911103301578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,1024,0.0069644442862934535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,8192,0.018100443813535903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,4096,0.025536000728607178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,128,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,10240,0.025416889124446448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,3072,0.0418817765182919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,65536,0.0605857769648234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,2048,0.005933333188295364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,7168,0.010058666268984476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,2560,0.021399999658266704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,256,0.002171555534005165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,5120,0.008414222134484185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,10240,0.017543110582563613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,1024,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,3584,0.019321777754359774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,12288,0.018202667435010273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,12288,0.01716711123784383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,2048,0.006679111056857639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,7168,0.017859554953045316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,7168,0.009287999735938178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,1024,0.006219555520349079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,16384,0.017883555756674874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,2560,0.004875555634498596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,4096,0.028026666906144884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,8192,0.017295999659432303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,6144,0.08425511254204644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,3072,0.021028444170951843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,1536,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,10240,0.03532888823085361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,128,0.005616888817813661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,6144,0.01884977850649092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,6144,0.011945777469211154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,768,0.017735110388861764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,1024,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,16384,0.019899555378490023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,7168,0.010406222608354356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,2560,0.006421333385838403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,12288,0.019469334019554984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,3072,0.007302222152551015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,3072,0.028304888142479792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,3072,0.012740444805887012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,3072,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,3072,0.004898666507667965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,512,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,512,0.012048000262843238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,768,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,10240,0.01941422290272183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,2048,0.016183111402723525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,768,0.006224888894293044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,5120,0.15381156073676214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,8192,0.009757333331637913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,3584,0.01145244472556644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,3584,0.016501333978441026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,16384,0.09959377845128377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,64,0.0038977778620190094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,256,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,32,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,4096,0.011023110813564725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,12288,0.2591368887159559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,7168,0.011718221836619906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,2048,0.00424888895617591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,512,0.014175110393100314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,64,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,512,0.006613333192136552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,16384,0.015141333142916361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,5120,0.12151200241512722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,12288,0.013447999954223633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,10240,0.06938311126497057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,4096,0.008296888735559251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,4096,0.009558222360081142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,32,0.006205333189831839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,32,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,5120,0.008616000413894653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,64,0.015184000134468079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,1024,0.014470222923490735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,16384,0.018184888694021437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,6144,0.008991111483838823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,5120,0.008295111358165741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,16384,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,65536,0.1810506714714898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,6144,0.009881777895821465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,2560,0.007173333730962541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,2048,0.00647288892004225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,3072,0.018241778016090393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,256,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,1024,0.005376889059940974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,16384,0.05013599991798401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,3584,0.018939556346999276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,1024,0.0034862222770849862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,3584,0.014833778142929077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,32,0.04359111189842224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,7168,0.046163555648591786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,16384,0.007090667055712805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,2560,0.0069324444565508105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,65536,0.12196532885233562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,3072,0.009563555320103964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,512,0.021908443835046556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,5120,0.015614221493403116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,12288,0.12679733170403376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,10240,0.017882666654056974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,768,0.002862222285734283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,7168,0.004927999857399198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,32,0.0018257778137922287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,1536,0.021312000022994146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,10240,0.016540444559521146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,1024,0.003549333247873518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,512,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,3072,0.01813066667980618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,16384,0.0365280012289683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,12288,0.018538667096032035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,7168,0.012272888587580787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,2560,0.02829155657026503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,2560,0.006736889067623351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,768,0.01648355523745219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,128,0.0019395556300878525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,3584,0.018215111560291715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,3072,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,65536,0.271616882748074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,1024,0.0188773340649075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,12288,0.014360000689824423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,768,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,768,0.008959111240175035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,3072,0.01887999971707662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,2560,0.008299555215570662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,12288,0.02402577797571818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,10240,0.11709866258833145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,256,0.002480000050531493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,8192,0.13948089546627468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,16384,0.44699732462565106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,512,0.0032417778339650896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,5120,0.009085333181752099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,2560,0.15225244892968073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,64,0.0018293333964215384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,16384,0.012902222573757172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,3072,0.1510817739698622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,2048,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,8192,0.017535110314687092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,2560,0.027043556173642475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,3584,0.007358222372002072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,7168,0.014178666803571912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,16384,0.02167822255028619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,2048,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,16384,0.10979555712805854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,2048,0.0042408886882993905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,2048,0.09279021951887344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,512,0.01310400002532535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,64,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,2048,0.015828443898095023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,65536,0.08271466361151801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,5120,0.07643110884560479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,1536,0.021306667062971327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,64,0.002198222196764416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,1536,0.006258666515350342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,128,0.027391110857327778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,512,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,65536,0.015392889579137167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,2048,0.007839110990365347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,16384,0.022615111536449854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,2560,0.016139555308553908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,64,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,2560,0.007311999797821045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,6144,0.01773155563407474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,64,0.01548266741964552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,2048,0.006950221955776215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,256,0.015533333023389181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,8192,0.02520177761713664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,65536,0.0731475551923116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,4096,0.03028800090154012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,3584,0.007594666547245449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,512,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,16384,0.0284551117155287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,3584,0.011037333144081963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,512,0.002568888788421949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,4096,0.011986666255527072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,8192,0.03096888793839349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,128,0.005061333378156026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,12288,0.021706667211320665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,256,0.014471999473041959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,128,0.002492444382773505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,65536,3.1308337317572703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,12288,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,10240,0.06571466392940946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,3072,0.007394666473070781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,6144,0.011298666397730509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,12288,0.053541332483291626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,128,0.00489688871635331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,1024,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,768,0.0032355555643637977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,3584,0.023631110787391663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,3072,0.007312888900438945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,1024,0.006237333433495627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,64,0.009705777797434065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,16384,0.04445777667893303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,6144,0.017533333765135873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,12288,0.016859556237856548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,1024,0.0028968888024489083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,12288,0.02232711182700263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,3072,0.004335111214054955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,1536,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,512,0.014496889379289416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,7168,0.033979554971059166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,10240,0.038536889685524836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,12288,0.014141332772043018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,512,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,65536,0.08153155777189466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,3072,0.007595555649863348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,4096,0.015805333852767944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,768,0.01719288859102461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,3072,0.01105955574247572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,256,0.010370666782061258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,2560,0.009701333112186855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,128,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,1024,0.0032151111712058387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,4096,0.04385511080423991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,128,0.00591822216908137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,7168,0.015160888433456421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,12288,0.01720622181892395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,8192,0.7247822019788953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,7168,0.025504888759719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,2560,0.007988444632954067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,8192,0.013853333062595792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,128,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,5120,0.012147555748621622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,2560,0.4532515737745497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,8192,0.012080888781282636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,1536,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,1024,0.0058853332367208265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,16384,1.1042391459147136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,1024,0.004316444612211651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,10240,0.021111999948819477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,2048,0.005044444567627377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,5120,0.01924888955222236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,6144,0.01202222208182017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,2048,0.022336888644430373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,5120,0.013909333282046847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,16384,0.04143111242188348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,10240,0.012793777717484368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,3072,0.014160000615649752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,12288,0.017888888716697693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,256,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,3584,0.12445955806308323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,6144,0.041434665520985924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,1536,0.005956444475385878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,8192,0.009411555197503831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,7168,0.2593475447760688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,7168,0.23814667595757377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,8192,0.014491554763582019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,3584,0.008988444175985124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,4096,0.06582399871614245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,256,0.0866862204339769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,1024,0.00589155571328269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,16384,0.02534488836924235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,6144,0.03254488772816128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,1024,0.01386577801571952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,2560,0.007289778027269576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,64,0.003843555433882607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,2560,0.021895999709765118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,128,0.012061333490742577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,65536,0.08393688996632893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,5120,0.017231111725171406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,7168,0.029919111066394385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,65536,0.22863645023769805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,65536,0.035727110173967146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,3584,0.007311111523045435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,3584,0.2440586619906955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,2048,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,12288,0.021543999512990315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,768,0.018019555343521964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,4096,0.00925688859489229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,8192,0.12143911255730523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,7168,0.007817777494589487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,128,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,768,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,16384,0.011071999867757162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,1024,0.0606222218937344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,128,0.005617777920431561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,16384,0.024400000770886738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,1536,0.009304000271691216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,64,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,3584,0.0074311114019817775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,2560,0.004584889031118817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,3584,0.019592000378502738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,65536,0.4998106426662869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,10240,0.020992888344658747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,10240,0.3522320058610704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,1024,0.0031742221779293488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,65536,0.0936008890469869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,128,0.004860444615284602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,8192,0.013722666435771518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,256,0.01722311145729489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,4096,0.008027555214034187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,12288,0.007537778053018782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,16384,0.05140533381038242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,3072,0.043419556485282056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,5120,0.10455466641320123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,3072,0.010819555156760745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,768,0.0031759999692440033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,2048,0.007376000285148621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,10240,0.04005600015322367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,65536,0.06272977590560913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,512,0.0053084443012873335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,256,0.009352888498041365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,65536,0.14623022079467773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,2048,0.015619556109110514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,8192,0.023077333966890972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,1024,0.014448889427714877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,12288,0.020484444167878892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,1536,0.005966222120655908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,65536,0.08416799704233806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,32,0.009695111049546136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,768,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,8192,0.02278311053911845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,1536,0.003527111063400904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,2560,0.02795377704832289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,16384,0.7433190875583224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,3584,0.0749537746111552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,8192,0.08665688832600911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,7168,0.11992088953653972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,3072,0.08868710862265693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,1024,0.02349511120054457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,5120,0.019132445255915325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,768,0.03437866767247518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,8192,0.0177502218219969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,10240,0.01719911065366533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,7168,0.0713955561319987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,3584,0.012866666747464074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,4096,0.021439111895031396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,6144,0.01689511040846507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,2560,0.006912888752089606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,32,0.006256000035338932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,768,0.002941333378354708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,1536,0.029071112473805744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,768,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,2048,0.006594666590293248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,128,0.002132444539003902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,3584,0.03266755408710904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,8192,0.019759999381171334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,768,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,3072,0.014159111513031853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,4096,0.012108444339699216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,65536,0.11363111601935492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,1024,0.009372444616423713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,12288,0.013761777844693927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,64,0.006288888967699475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,8192,0.014366222752465142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,1024,0.013591999808947245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,5120,0.004868444469239977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,6144,0.032799111472235784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,3072,0.02689244515366024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,6144,0.009651555783218807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,7168,0.02941511074701945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,2560,0.05333422289954292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,10240,0.011730666789743634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,8192,0.032524443335003324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,5120,0.030198223061031763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,12288,0.014562666416168213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,64,0.002464888824356927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,5120,0.008656889200210571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,16384,0.01904533306757609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,768,0.0028328889360030494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,32,0.002179555594921112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,12288,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,3584,0.0655848913722568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,1536,0.02699733277161916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,64,0.001822222231162919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,7168,0.034146666526794434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,12288,0.01978400018480089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,512,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,512,0.0041804446114434134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,3072,0.011095110740926532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,7168,0.026895110805829365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,768,0.009206222163306342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,16384,0.021924444370799597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,10240,0.12840622001224095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,2048,0.016501333978441026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,1024,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,3584,0.007986666427718269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,7168,0.018896889355447557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,65536,0.08458577924304539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,3584,0.10824444558885361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,768,0.006643555644485686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,4096,0.012711111042234631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,3072,0.004584889031118817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,768,0.02335022224320306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,256,0.011651555697123209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,3584,0.0074755557709270054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,3072,0.017619555195172627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,4096,0.00926577796538671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,64,0.16276710563235813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,256,0.0031893334041039148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,12288,0.01682933337158627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,2048,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,12288,0.10220977995130752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,6144,0.009328888522254096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,10240,0.02644355595111847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,256,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,32,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,2560,0.01790577835506863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,10240,0.016678222351604037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,3584,0.006585777633719974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,256,0.00553688903649648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,2560,0.006622222148709827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,12288,0.05123466584417555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,2560,0.01483022173245748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,10240,0.031424889961878456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,16384,0.052320887645085655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,32,0.005528000079923206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,128,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,12288,0.022126222650210064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,16384,0.004551110996140374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,10240,0.020568889048364427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,128,0.013078221844302284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,8192,0.014353776971499125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,2048,0.04032889008522034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,3584,0.017231111725171406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,768,0.013443555268976422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,512,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,256,0.003118222165438864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,5120,0.013482666677898832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,1536,0.02386844489309523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,1024,0.0031235555393828284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,3072,0.006984000404675801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,64,0.006605333338181178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,10240,0.029955555995305378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,3584,0.007963555554548899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,256,0.01750666730933719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,3072,0.024384000235133704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,768,0.016795555750528973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,3072,0.0076044441925154785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,10240,0.0503422220547994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,3072,0.02252799934811062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,2048,0.005024888863166173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,16384,0.027318222655190363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,512,0.005995555470387141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,128,0.0022257777551809945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,10240,0.01642400026321411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,2560,0.01477688882086012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,3584,0.022221333450741235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,8192,0.05054222212897407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,768,0.006291555447710886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,512,0.006646222124497096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,1536,0.005935110979610019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,5120,0.016502221425374348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,1024,0.007623111208279927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,128,0.006973333656787872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,768,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,512,0.03490044342146979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,4096,0.03161155515246921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,16384,0.04113422168625726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,3072,0.006966222491529252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,64,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,3584,0.5489368968539768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,12288,0.07879288991292317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,1024,0.006955555743641323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,512,0.04484533270200094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,7168,0.019713777634832595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,12288,0.02844088938501146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,2560,0.02437422176202138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,128,0.01871466636657715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,2048,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,65536,0.16399644480811224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,4096,0.019925332731670804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,3584,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,10240,0.015000889698664347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,10240,0.021073778470357258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,512,0.0024746666765875286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,3584,0.02780888809098138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,64,0.004873777843183941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,3584,0.023159111539522808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,7168,0.01721600029203627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,3584,0.007505777809354994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,12288,0.018065777089860704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,1536,0.009316444396972656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,2048,0.04783111148410373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,512,0.014147554834683737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,6144,0.010022222167915767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,8192,0.013853333062595792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,3072,0.010007999837398529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,16384,0.021915555000305176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,16384,0.012818666795889536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,128,0.014153778553009033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,10240,0.01787644459141625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,128,0.014495111174053617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,4096,0.007916444705592262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,8192,0.009750222166379293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,7168,0.02201333310869005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,16384,0.03182755576239692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,768,0.017124444246292114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,1536,0.01444355481200748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,7168,0.009676444033781687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,128,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,65536,0.1947573290930854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,10240,0.012523555921183692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,7168,0.031014220582114324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,32,0.002179555594921112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,6144,0.005946666830115848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,128,0.010034666293197209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,1536,0.004808889081080754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,5120,0.011727111207114326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,5120,0.02513422237502204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,2560,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,64,0.005685333162546158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,128,0.013459555804729462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,64,0.0018177777528762817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,10240,0.012714666624863943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,512,0.04455911119778951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,128,0.002159111201763153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,8192,0.3564346631368001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,3072,0.009024000002278222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,256,0.009692444569534725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,12288,0.01811555524667104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,2048,0.006919999917348226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,2048,0.015459555718633862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,3584,0.016514667206340365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,12288,0.06688888867696126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,4096,0.007622222105662028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,32,0.0025191110455327565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,32,0.015782222151756287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,2560,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,3584,0.021385777327749465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,10240,0.017090666625234816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,3584,0.030156443516413372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,12288,0.018241778016090393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,512,0.010693333215183683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,10240,0.04615644282764859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,16384,0.30406665802001953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,7168,0.01773955590195126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,6144,0.00942844483587477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,8192,0.0052844443255000645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,1536,0.028720001379648846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,65536,6.470341152615017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,2048,0.006351999938488007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,2560,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,128,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,2560,0.006816888848940532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,7168,0.02366311185889774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,32,0.0031866667171319327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,32,0.006583111153708563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,10240,0.006143110907740063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,2560,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,3584,0.020602666669421725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,512,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,768,0.009556444154845344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,512,0.003866666720973121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,4096,0.006544000158707301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,3584,0.010528888967302112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,1024,0.005892444401979446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,16384,0.020195555355813768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,256,0.014499555031458536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,3072,0.348743995030721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,1536,0.006148444281684027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,1536,0.020918221937285528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,512,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,5120,0.13082310888502333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,8192,0.05637333128187391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,256,0.015853333804342482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,128,0.027486221657858953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,5120,0.011319110790888468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,2560,0.005010666532648934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,5120,0.11481066544850667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,4096,0.012432888978057437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,7168,0.01752000053723653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,10240,0.03242399957444932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,1024,0.006268444574541516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,64,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,10240,0.0428435570663876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,256,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,12288,0.0343413319852617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,65536,0.24844622611999512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,12288,0.01999022232161628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,3584,0.01920088960064782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,2048,0.006281777802440856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,5120,0.10444444417953491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,16384,0.06084444125493368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,3584,0.007256000406212277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,2560,0.01644088824590047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,1024,0.002871111035346985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,1024,0.007517333659860823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,5120,0.024499555428822834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,10240,0.01664177742269304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,2560,0.03307022319899665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,3584,0.09765422344207764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,6144,0.06322222285800509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,10240,0.06481422318352593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,4096,0.02502577834659153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,128,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,1536,0.0038986665507157645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,1024,0.004600000050332812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,2048,0.1928622192806668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,65536,0.0769137806362576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,2560,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,2048,0.021755556265513103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,3072,0.040469331873787776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,65536,0.02383022175894843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,3072,0.010441777606805166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,8192,0.07440977626376681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,4096,0.024528000089857314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,1536,0.006584888945023219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,3072,0.017076444294717576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,256,0.005937777873542573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,256,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,5120,0.008646222452322641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,12288,0.262588447994656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,1024,0.007450666692521837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,2560,0.0069520001610120135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,2048,0.011408000356621213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,128,0.002159111201763153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,8192,0.01443911095460256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,256,0.009162666896979014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,256,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,3584,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,7168,0.04550133479966057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,3584,0.011384000380833944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,65536,0.07984266678492229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,16384,0.04304888844490051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,5120,0.0351217786471049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,3584,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,3072,0.006842666616042455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,512,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,128,0.01347733371787601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,2560,0.0045422220395671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,12288,0.03362755642996894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,65536,0.05635288688871595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,32,0.003863111138343811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,2048,0.007640889121426477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,8192,0.01511111193233066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,512,0.004993777722120285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,5120,0.019087110956509907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,3584,0.007747555772463481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,4096,0.008761777646011775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,32,0.006027555714050929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,7168,0.012669333153300814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,10240,0.014906666345066495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,128,0.006815111057625876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,4096,0.008117333054542542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,2048,0.05643555853101942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,3584,0.0218248897128635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,2048,0.019239111079110038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,3072,0.009200889203283522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,16384,0.5687706735399034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,2560,0.10280266735288833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,16384,0.027447111076778833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,768,0.0029271110478374693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,768,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,1536,0.020240000552601285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,8192,0.022592888938056097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,16384,0.040884445110956825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,65536,0.04684622089068095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,3072,0.006908444480763541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,7168,0.026710222164789837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,3072,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,7168,0.03533333208825853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,512,0.012457778056462606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,1024,0.013438222308953604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,16384,0.01682400041156345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,8192,0.03657689028316074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,6144,0.0048942222363419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,4096,0.011559111376603445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,768,0.0028311111446883944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,8192,0.0099653336736891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,10240,0.036158220635520086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,16384,0.043403555949529014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,3072,0.007718222008811102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,2048,0.15275822745429143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,8192,0.04124177826775445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,7168,0.02630222174856398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,3072,0.008613333106040955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,256,0.013751111096805997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,64,0.002833777831660377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,12288,0.013771555489963956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,768,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,2048,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,16384,0.0927084419462416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,128,0.007981333467695449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,128,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,256,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,10240,0.02158311174975501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,1024,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,1024,0.01651022169325087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,512,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,2560,0.009583999713261923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,4096,0.02865155537923177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,16384,0.13987378279368082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,256,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,32,0.0021884443445338146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,256,0.00829777783817715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,32,0.003173333282272021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,8192,0.01513066722287072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,512,0.01961955593691932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,16384,0.07111199696858723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,65536,0.08347022533416748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,10240,0.02520977788501316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,12288,0.01223111069864697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,10240,0.005379555539952383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,12288,0.058628446525997586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,128,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,6144,0.02093244426780277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,768,0.0026462222966882917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,128,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,3584,0.011201777391963534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,2560,0.00902844468752543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,7168,0.009669333696365356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,1024,0.01716711123784383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,3584,0.007438222567240397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,2048,0.006270222365856171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,512,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,32,0.010390222072601318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,3584,0.013077333569526672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,12288,0.01870755520131853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,65536,0.24192889531453451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,64,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,10240,0.03755377729733785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,3072,0.00434933313065105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,768,0.0052879999081293745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,10240,0.029538667864269678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,16384,0.13939200507269964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,10240,0.011919111013412476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,2560,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,512,0.008920888933870528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,3072,0.009843555589516958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,768,0.013986666997273764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,256,0.005542222410440445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,16384,0.16418577565087214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,768,0.0028248888750871024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,8192,0.018294221825069852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,16384,0.015840889679061044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,128,0.012811555630630918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,768,0.007364444434642792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,1024,0.0032142222755485107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,7168,0.17937421798706055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,2560,0.015496888094478183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,2048,0.006796444455782573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,1536,0.0041706665522522396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,3072,0.007284444239404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,2048,0.024195555183622573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,10240,0.015056888262430826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,12288,0.009669333696365356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,3584,0.007605333295133378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,65536,0.3735617796579997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,3584,0.04597066508399116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,12288,0.036213331752353244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,5120,0.01071111112833023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,5120,0.004931555440028508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,128,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,1536,0.021107556091414556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,12288,0.252674659093221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,256,0.01479555500878228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,7168,0.009654222263230218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,1024,0.0028328889360030494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,1536,0.028809779220157202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,10240,0.05031822125116984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,256,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,1024,0.0053635554181204895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,512,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,10240,0.039874666266971164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,1536,0.007669332954618666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,4096,0.010012444522645738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,768,0.01870933340655433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,3072,0.007249777515729268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,16384,0.02751200066672431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,3584,0.008002666963471307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,10240,0.05964622232649061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,32,0.005192000004980299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,6144,0.009375111096435124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,64,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,16384,0.07574311229917738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,7168,0.013030221892727746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,5120,0.020452444752057392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,16384,0.00814666681819492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,10240,0.019899555378490023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,6144,0.017601778109868366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,2048,0.020613332589467365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,3072,0.01109688894616233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,3584,0.007633777956167857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,12288,0.053655998574362866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,256,0.011678222152921887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,2560,0.006695999867386288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,10240,0.12137777275509304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,2560,0.004875555634498596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,256,0.006568000134494569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,256,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,32,0.006240889016124938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,10240,0.045048889186647206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,512,0.00488711107108328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,12288,0.10385777552922566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,1536,0.005596444424655702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,5120,0.00831200016869439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,1536,0.021126222279336717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,3584,0.009838222629494136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,128,0.0196142229768965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,10240,0.1087262233098348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,256,0.0027804445061418745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,4096,0.4517679744296604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,64,0.023319111929999456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,1024,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,3072,0.015776000089115567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,2560,0.02721066607369317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,12288,0.03480089041921828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,1536,0.00832622249921163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,512,0.009677333136399588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,2560,0.007931555310885111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,16384,0.04334933227962918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,65536,0.14948799875047472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,12288,0.036531554328070745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,768,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,1024,0.021351110604074266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,65536,0.024681778417693243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,512,0.002895111011134254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,3072,0.025136000580257837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,3584,0.015820443630218506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,512,0.005541333307822545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,256,0.015473778049151102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,8192,0.019853333632151287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,768,0.005340444544951121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,512,0.010682666467295753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,512,0.04225066635343763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,2560,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,512,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,8192,0.015654222832785714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,16384,0.01632355484697554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,3584,0.017161778277821012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,1024,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,2560,0.02232622272438473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,16384,0.7058364550272623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,10240,0.04495199852519565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,1024,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,2048,0.007790221936172909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,3072,0.14254755443996853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,32,0.0021280000607172647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,128,0.005631111148330901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,512,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,6144,0.018093332648277283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,256,0.003195555466744635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,128,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,4096,0.00794400026400884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,3584,0.041843556695514254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,3584,0.011716444459226398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,768,0.0031759999692440033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,2048,0.0879431102010939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,1536,0.008293333152929941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,768,0.006301333506902059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,32,0.0051893335249688895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,3072,0.004862222406599256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,128,0.0035208890007601846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,128,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,65536,0.08717066711849636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,6144,0.046182221836513944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,512,0.005377777748637729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,3584,0.01853511068556044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,65536,0.09842311011420356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,64,0.005945777727497949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,4096,0.014813333749771118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,16384,0.018198221921920776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,4096,0.010414222048388587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,128,0.004937777916590373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,12288,0.015722667177518208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,6144,0.008945778012275696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,256,0.01312444441848331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,2048,0.01477777792347802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,3072,0.03187555405828688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,8192,0.02093244426780277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,5120,0.010447111394670276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,5120,0.015809777710172866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,8192,0.046912001238928906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,10240,0.005919111271699269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,128,0.014856000741322836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,128,0.003887999802827835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,2048,0.005056000004212062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,4096,2.64449331495497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,65536,0.16635466946495905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,3584,0.035952889257007174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,256,0.0028008888992998335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,768,0.003947555604908201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,1536,0.004571555389298333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,1536,0.07895021968417697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,64,0.0063635553750726916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,2048,0.01882133384545644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,12288,0.018605333235528734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,256,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,7168,0.8145333396063911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,512,0.013811555173661975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,768,0.006607111129495833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,2048,0.00848355558183458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,65536,0.06804355647828844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,1536,0.009707555174827576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,2560,0.009992000129487779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,65536,0.02566844390498267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,2048,0.0038577777643998465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,5120,0.03560977843072679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,4096,0.007657777931955125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,10240,0.022005332840813532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,2048,0.006224888894293044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,6144,0.010673777924643623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,3072,0.04712444543838501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,128,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,32,0.00179822225537565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,4096,0.008990222381220924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,6144,0.1313528882132636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,65536,0.9182035658094617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,256,0.012110222544935016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,2560,0.016905777984195285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,4096,0.006121777825885349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,6144,0.009503111243247986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,7168,0.05363288852903578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,512,0.0024746666765875286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,512,0.01809599995613098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,65536,0.14233956072065565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,4096,0.008303999900817871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,3072,0.006955555743641323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,16384,0.025221332907676697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,16384,0.02160266704029507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,768,0.0191600008143319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,32,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,3072,0.01035822182893753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,5120,0.03740088807211982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,16384,0.006631999793979857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,1024,0.02014222244421641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,6144,0.009292444421185387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,256,0.008023999631404877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,8192,0.08091377549701266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,2048,0.006309333360857434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,1024,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,8192,0.046499556965298124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,5120,0.0127688886390792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,3584,0.009324444664849175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,32,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,16384,0.02404266595840454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,8192,0.04340977801216973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,8192,0.021894221504529316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,768,0.00562755556570159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,3584,0.005643555687533484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,6144,0.0255039996571011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,768,0.005512889060709212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,2048,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,7168,0.04142755601141188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,1024,0.008674666285514832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,7168,0.017493334081437852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,2560,0.01888799998495314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,256,0.005886222339338726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,10240,0.01975288821591271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,128,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,512,0.00407644444041782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,3072,1.0889111624823677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,65536,0.02199111051029629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,16384,0.039423998859193586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,768,0.005951111101441913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,2560,0.0347724457581838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,2048,0.015170666906568738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,512,0.008277333445019191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,8192,0.02507200009293026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,128,0.061894224749671094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,8192,0.0052328887912962176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,32,0.0028693332440323303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,7168,0.009993777506881291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,6144,0.01623288955953386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,10240,0.01960088809331258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,12288,0.0340320004357232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,768,0.0028293333533737394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,1024,0.006265777680608962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,3584,0.007332444190979004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,3072,0.007293333609898885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,3072,0.01753777762254079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,12288,0.071651558081309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,768,0.01556711064444648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,128,0.009597332941161262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,32,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,2560,0.08040177822113037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,6144,0.025401777691311304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,16384,0.015782222151756287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,8192,0.011722666521867117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,8192,0.012135110795497894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,2048,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,2048,0.00629688882165485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,8192,0.01608533329433865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,65536,0.08338577879799737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,2048,0.006803555621041193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,7168,0.044567998912599355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,3584,0.007320889168315464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,12288,0.024751999311976965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,7168,0.016358221570650738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,12288,0.04697511262363858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,10240,0.15209955639309353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,2560,0.006907555378145642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,65536,0.17975466781192354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,3072,0.013155555559529198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,768,0.007300444775157505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,128,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,32,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,5120,0.011731555892361535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,3072,0.007986666427718269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,2560,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,5120,0.0181368887424469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,128,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,256,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,8192,0.30196621682908803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,4096,0.01722755531469981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,3584,0.014430221584108142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,128,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,512,0.007592889169851939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,7168,0.019911999503771465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,16384,0.03605688942803277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,768,0.005640000104904175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,4096,0.03868800070550706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,10240,0.04931111137072245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,64,0.001847111102607515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,7168,0.16167822149064806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,12288,0.009690666364298927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,10240,0.03110044532352024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,8192,0.01959466603067186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,512,0.0041866666740841335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,5120,0.05137066708670723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,3072,0.06625688738293119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,256,0.002471111093958219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,3072,0.034303999609417386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,7168,0.049866666396458946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,6144,0.010622222390439777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,4096,0.013750221994188098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,2048,0.004461333155632019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,10240,0.3154088921017117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,10240,0.02780088782310486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,4096,0.01611022154490153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,6144,0.010238221950001186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,64,0.0018133332745896445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,1536,0.011591999895042844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,5120,0.025976889663272437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,12288,0.00961688905954361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,16384,0.016943999462657504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,2560,0.0073315559162033936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,65536,0.02422044508987003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,7168,0.00700711127784517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,1536,0.006298666612969504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,10240,0.01036977767944336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,3584,0.04314755400021871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,5120,0.07167110840479533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,12288,0.5018373595343696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,2048,0.0181342214345932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,2560,0.00453599997692638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,32,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,12288,0.045327110422982105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,2048,0.009309333231714036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,256,0.018336888816621568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,16384,0.015556444724400839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,7168,0.007296000089910295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,8192,0.012117333710193634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,7168,0.048882666561338634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,7168,0.03972444600529141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,2560,0.006573333508438534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,32,0.0025235555238193935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,2048,0.00831288927131229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,8192,0.08510844575034247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,256,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,768,0.008639111287064022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,1024,0.0053386667536364655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,32,0.0021128888345426982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,2560,0.016233777006467182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,2560,0.01088355564408832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,10240,0.01625688870747884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,64,0.002838222309947014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,12288,0.015240889456537036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,8192,0.014550222290886773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,512,0.002508444504605399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,8192,0.019917334119478863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,1024,0.003530666646030214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,768,0.0052826665341854095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,12288,2.029327180650499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,3072,0.010121777653694153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,32,0.010023111270533668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,65536,0.21786044703589547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,32,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,16384,0.4602035416497125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,3584,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,7168,1.0862337748209636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,12288,0.03126488791571723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,8192,0.013733333183659447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,256,0.002536888958679305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,6144,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,65536,0.062223109934065074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,128,0.004929777648713854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,128,0.0021760000122918025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,32,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,7168,0.01276533305644989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,512,0.012764444781674279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,1536,0.005945777727497949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,128,0.00254133323000537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,10240,0.017491555876202054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,1536,0.006255110932721033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,1536,0.06641511122385661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,64,0.006230222268237009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,256,0.007910221815109253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,128,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,256,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,4096,0.007947555846638149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,65536,0.07546844747331408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,10240,0.020202666521072388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,1024,0.005961777849329843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,3584,0.015153777268197803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,1536,0.004204444587230682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,64,0.012404444317022959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,768,0.02373955481582218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,8192,0.017453332742055256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,16384,0.014351111319330005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,128,0.006248000181383557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,128,0.002135111019015312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,768,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,2048,0.022190221481853064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,6144,0.01310755560795466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,512,0.003955555458863576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,4096,0.016520000166363187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,8192,0.011043555206722684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,5120,0.008298666940795051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,1536,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,1536,0.006573333508438534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,256,0.004936000125275718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,2048,0.006569777925809224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,64,0.0021173333128293357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,32,0.0018524444765514797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,1024,0.003474666840500302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,3584,0.015928000211715698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,6144,0.13505600558386907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,5120,0.10512355301115249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,7168,0.03626399901178148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,512,0.006193777753247156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,3072,0.01910666624704997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,768,0.005617777920431561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,65536,0.15573599603441027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,6144,0.012551999754375882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,65536,0.22066311041514078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,5120,0.01719999975628323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,128,0.006092444476154115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,4096,0.015615999698638916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,5120,0.011429333024554782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,8192,0.023367111881573994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,3072,0.02177511155605316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,3584,0.007273777491516537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,128,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,8192,0.011027555498811932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,1536,0.014509333504570855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,12288,0.8905546400282117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,6144,0.012141333685980903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,8192,0.03747466537687514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,5120,0.011744889120260874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,8192,0.01794399983353085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,256,0.0052328887912962176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,1024,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,1024,0.026054221722814772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,256,0.005943111247486538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,768,0.02998222245110406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,6144,0.009313777916961247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,10240,0.21845600340101454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,128,0.018555555078718398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,8192,0.011949333051840464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,256,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,2048,0.08515999714533488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,8192,0.015484443969196744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,65536,0.22232355011834037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,4096,0.0720613333914015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,3072,0.02943999899758233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,3584,0.007319110963079665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,256,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,512,0.002483555633160803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,16384,0.03439733386039734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,3072,0.007279111279381647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,2048,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,2560,0.012109333442317115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,12288,0.08679288625717163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,16384,0.07565066549513075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,4096,0.04275733232498169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,512,0.005249777601824866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,5120,0.011377777490350934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,4096,0.027853333287768896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,3072,0.010613333019945355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,32,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,512,0.005252444495757421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,3072,0.004559999952713649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,10240,0.037436445554097496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,5120,0.029860443539089624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,2048,0.009312888814343346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,1024,0.09474311272303264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,4096,0.07715466949674818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,16384,0.0236462222205268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,12288,0.05184355709287855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,3072,0.008709333009190029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,32,0.002017777827050951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,1024,0.0029724445194005966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,256,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,128,0.014078223043017916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,768,0.00563466673096021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,256,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,1024,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,128,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,5120,0.009377777576446533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,4096,0.10868355300691392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,5120,0.02513066596455044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,1024,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,768,0.003939555750952827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,768,0.008527999950779809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,10240,0.012080888781282636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,65536,0.38919112417433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,512,0.021331555313534204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,16384,0.04106222258673774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,128,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,4096,0.024679111109839544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,64,0.0018444444156355327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,1536,0.006940444310506185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,64,0.003887999802827835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,512,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,1024,0.008738666772842407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,1536,0.005945777727497949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,7168,0.06312266985575359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,16384,0.021792888641357422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,512,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,12288,0.02100444502300686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,256,0.00554666668176651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,3584,0.03229511115286086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,65536,0.464580429924859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,768,0.016539555456903245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,32,0.00871288859181934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,3584,0.017118222183651395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,512,0.004218666503826777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,65536,0.08040799697240193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,1536,0.006005333529578314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,4096,0.016179554992251925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,512,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,768,0.024506666594081458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,3072,0.007282666862010956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,1024,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,1024,0.003583111282851961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,16384,0.0625457763671875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,7168,0.009345778160625035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,7168,0.019202666150199044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,1536,0.032080001301235624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,3072,0.00453244439429707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,4096,0.07508533530765109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,5120,0.03669955664210849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,3584,0.008418666819731394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,7168,0.01682933337158627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,4096,0.0038888889054457345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,4096,0.014512888259357877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,1024,0.006272000157170826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,512,0.06392533249325223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,256,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,2048,0.006265777680608962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,1536,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,1536,0.006222222414281633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,3072,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,12288,0.012712000144852532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,128,0.01443911095460256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,128,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,128,0.006234666539563074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,1024,0.005610666755172942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,768,0.005967999911970562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,2560,0.07830222447713216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,128,0.013074666261672974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,7168,0.00488800017370118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,4096,0.015192889504962497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,2048,0.011538666983445486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,16384,0.02601066728432973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,128,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,32,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,64,0.0028071111689011254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,1024,0.005922666854328579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,768,0.003208000005947219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,1536,0.006065777606434292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,768,0.09371555513805813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,768,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,256,0.004556444370084339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,64,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,128,0.01610311037964291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,1536,0.017535110314687092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,10240,0.04521422253714668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,8192,0.01443555619981554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,3584,0.008649778034951951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,3584,0.018173333671357896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,8192,0.161789337793986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,2560,0.006801777829726537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,7168,0.022351110974947613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,16384,0.01613333324591319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,8192,0.009992000129487779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,8192,0.032456000645955406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,12288,0.004941333499219683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,65536,0.21760266356998018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,128,0.016181333197487723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,3072,0.023568888505299885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,128,0.004960000101062986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,65536,0.08054222000969781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,7168,0.010371555884679159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,3584,0.02106488909986284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,128,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,256,0.004914666629499859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,256,0.018182223041852314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,1536,0.006292444550328785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,3584,0.01605955594115787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,1536,0.003825777934657203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,1024,0.012077333198653327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,5120,0.00830666638082928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,6144,0.00869599978129069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,6144,0.010942222343550788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,7168,0.017529777354664274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,5120,0.023022222850057814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,768,0.01867377758026123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,2048,0.018936889039145578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,768,0.006005333529578314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,64,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,1536,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,1024,0.03790844480196635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,3072,0.00720177756415473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,1536,0.0038248888320393036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,256,0.00254133323000537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,7168,0.021205332544114854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,16384,0.052461332745022245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,16384,0.03509955604871114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,2560,0.013336000343163809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,1536,0.016198222835858662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,2048,0.018518222702874076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,512,0.010199111368921068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,65536,7.505174424913194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,4096,0.012840000291665396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,65536,1.4537199868096247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,1024,0.004583111239804162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,512,0.008271999657154083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,256,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,1536,0.006100444330109491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,768,0.0029004443850782183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,128,0.002226666650838322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,768,0.002961777771512667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,2560,0.01616533266173469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,1536,0.005985777825117111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,32,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,65536,0.3126595550113254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,16384,0.3328639931148953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,4096,0.007947555846638149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,1536,0.007186666958861881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,32,0.001828444500764211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,256,0.004915555732117759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,65536,0.32250134150187176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,10240,0.0930631094508701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,5120,0.02795377704832289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,1536,0.017504889104101393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,16384,0.016833777228991192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,128,0.014112000664075216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,4096,0.024157333705160353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,1024,0.026945778065257605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,5120,0.014327110515700446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,6144,0.030504888958401147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,16384,0.12292444705963135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,1024,0.02923555506600274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,8192,0.02400355537732442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,512,0.06100533405939738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,12288,0.05239555570814344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,8192,0.015792888071801927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,4096,0.008294222255547842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,32,0.006559111177921295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,128,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,10240,0.013768889009952545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,10240,0.05009955498907301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,64,0.007322666545708974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,16384,0.0904817779858907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,5120,0.00902844468752543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,1536,0.006227555374304454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,16384,0.019666666785875957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,512,0.04478044311205546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,16384,0.06566044357087877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,3072,0.060890667968326144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,7168,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,5120,0.015802666544914246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,10240,0.014487110906177096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,32,0.00281866660548581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,256,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,2048,0.017117333081033494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,3072,0.04566488994492424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,768,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,5120,0.02327022287580702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,256,0.002546666603949335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,10240,0.5873653093973795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,10240,0.005944889038801193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,3584,0.009485333330101436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,5120,0.024128889044125874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,2048,0.011519110865063138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,8192,0.01108088923825158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,256,0.0028400001012616684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,2048,0.01274400038851632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,256,0.0021840000732077495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,64,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,4096,0.010732444624106089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,12288,0.018561777141359117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,4096,0.09079821904500325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,5120,0.015543111496501498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,16384,0.05729777945412529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,3584,0.006207110981146495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,128,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,8192,0.027856889698240492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,5120,0.13331911298963758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,256,0.014452444182501899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,65536,0.13519733481936985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,128,0.014159111513031853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,4096,0.004543111142185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,768,0.0069431112044387394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,1024,0.004601777841647466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,7168,0.010011555420027839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,768,0.005370666583379109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,3584,0.01033866653839747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,16384,0.04821599854363335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,2560,0.012952889005343119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,1024,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,7168,0.009653333160612319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,5120,0.011630222201347351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,2048,0.014402666025691561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,5120,0.06079022089640299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,65536,0.13938489225175646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,3584,0.046588444047504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,3584,0.02537422213289473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,128,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,1024,0.0031946665710873077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,2048,0.01993333299954732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,1536,0.01510933372709486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,2560,0.013771555489963956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,1536,0.08279999759462145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,3072,0.03970311085383097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,512,0.015882665912310284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,16384,0.004920000003443824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,2048,0.006487999939256244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,16384,0.01643822259373135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,16384,0.023345778385798138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,7168,0.01662133302953508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,3584,0.009184888667530483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,3584,0.037315554089016385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,7168,0.009712888962692684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,16384,0.01759999990463257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,256,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,2560,0.03360177742110358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,10240,0.01204355557759603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,3584,0.027391110857327778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,768,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,3072,0.01072799993885888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,6144,0.07363288932376437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,1536,0.006311999840868845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,2560,0.009211555951171452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,3584,0.02933244572745429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,256,0.014487110906177096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,768,0.006908444480763541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,256,0.01686311099264357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,4096,0.04119111100832621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,3072,0.022631110416518316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,10240,0.011364444262451597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,128,0.008008889026112026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,256,0.007831110722488826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,32,0.002507555608948072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,3584,0.02385422256257799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,1024,0.02348711093266805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,4096,0.023750222391552393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,1536,0.005015111217896144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,256,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,2560,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,2560,0.006187555690606435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,12288,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,8192,0.011035555766688453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,65536,0.08798399898740981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,7168,0.01275022245115704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,2560,0.01519377695189582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,2048,0.006694222076071634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,3072,0.017498667041460674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,2048,0.006912888752089606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,3584,0.049597332874933876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,768,0.022985777921146814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,4096,0.10937244362301296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,3584,0.007174222005738153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,7168,0.022461333208613928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,2560,0.008759111166000366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,6144,0.04029688901371426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,2560,0.007259555160999298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,16384,0.01717155509524875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,7168,0.015304888288180033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,10240,0.011432888607184092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,1024,0.033746666378445096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,1024,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,65536,0.08335555262035793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,65536,2.242255104912652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,65536,0.067358222272661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,1536,0.007273777491516537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,10240,0.015824889143308003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,3584,0.007648000286685095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,128,0.006577777779764599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,2560,0.010979555547237396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,256,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,1024,0.0058711109062035876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,12288,0.004530666602982415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,5120,0.01741066740618812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,32,0.0021066665649414062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,6144,0.02803200152185228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,5120,0.01387555566098955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,3072,0.06172888808780246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,10240,0.017867555220921833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,16384,0.07039555576112536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,5120,0.012243555651770698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,2560,0.006955555743641323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,16384,0.03238666719860501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,4096,0.00834222220712238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,3584,0.025416889124446448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,6144,0.020118221640586853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,3584,0.04996888836224874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,6144,0.006240889016124938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,65536,0.31450843811035156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,16384,0.0659102201461792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,2560,0.009168888959619734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,128,0.002528000002106031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,512,0.014508444401952954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,1024,0.013459555804729462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,2560,0.014436443646748861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,256,0.012452444268597497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,2560,0.016964443855815463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,64,0.002542222125662698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,256,0.005984000033802456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,5120,0.01826755536927117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,768,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,3072,0.007615110940403408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,10240,0.07491822375191583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,768,0.004928888960017098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,32,0.0018533333722088074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,1536,0.006614222294754452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,3072,0.009384888741705153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,10240,0.041950222518708974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,6144,0.012092444631788464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,2048,0.020308444897333782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,4096,0.024338665935728285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,128,0.009089777866999308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,1536,0.005952000204059813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,256,0.00674044465025266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,512,0.019675556156370375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,16384,0.02233155568440755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,256,0.016950221525298226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,3072,0.009648000200589498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,2048,0.01571822166442871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,8192,0.016018667154841952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,3584,0.007341333561473423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,3584,0.1698533296585083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,10240,0.02759733299414317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,4096,0.00793599999613232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,32,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,65536,0.10368266370561387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,16384,0.16265511512756348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,1024,0.006254222244024277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,12288,0.017165333032608032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,128,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,2560,0.01475911173555586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,768,0.011167111496130625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,2560,0.013763555222087435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,2048,0.01994044416480594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,6144,0.09467289182874893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,1536,0.02439466615517934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,3072,0.016944888565275405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,32,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,2560,0.009648888475365108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,8192,0.005888000130653381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,512,0.015164444843928019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,4096,0.053744888967937894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,512,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,6144,0.03416000141037835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,256,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,2048,0.01761155492729611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,5120,0.028776887390348647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,16384,0.03956177830696106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,32,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,64,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,1536,0.020005333754751418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,7168,0.009764444496896531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,32,0.006572444405820634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,3072,0.020729778541458976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,3072,0.010683555569913654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,8192,0.055383112695482045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,16384,0.026000888811217412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,2560,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,3584,0.01035644445154402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,5120,0.036870221296946205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,4096,0.018385777870814007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,768,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,7168,0.057553774780697294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,1536,0.007161777880456712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,3584,0.008316444853941599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,1024,0.003170666595300039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,6144,0.022109333011839125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,4096,0.00813066628244188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,65536,0.015847111741701763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,128,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,12288,0.036469333701663546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,256,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,256,0.00655733338660664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,12288,0.08174310790167914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,128,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,12288,0.14850933021969265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,8192,0.0151573336786694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,6144,0.009564444422721863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,8192,0.0214373336897956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,6144,0.019945777124828763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,10240,0.03128800127241347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,10240,0.012490666574902005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,64,0.006632888896597757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,4096,0.008055110772450766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,768,0.005620444400442972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,16384,0.033921778202056885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,4096,0.02861422300338745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,4096,0.010809777511490716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,1536,0.006335111127959357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,512,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,64,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,1024,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,7168,0.010055111514197456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,8192,0.05279022124078539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,6144,0.047727111313078135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,3072,0.17632622188991973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,8192,0.016827555166350473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,7168,0.010607110957304636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,12288,0.012848888834317526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,65536,0.26518045531378853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,128,0.014490667316648694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,10240,0.024578667349285547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,3584,0.07131110959582858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,7168,0.0151591118839052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,1536,0.015154666370815702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,2560,0.006607999818192587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,2560,0.026407111022207472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,2560,0.014463111758232117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,128,0.014169777433077494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,3072,0.008392888638708327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,128,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,2048,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,1024,0.0031822222388452957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,2560,0.020346666375796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,4096,0.008068444828192392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,32,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,7168,0.5262231296963161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,64,0.006968888971540663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,8192,0.046647999021742076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,32,0.002838222309947014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,128,0.008959111240175035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,2048,0.006439111298984951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,3072,0.006530666516886816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,6144,0.11417955822414821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,1536,0.006265777680608962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,65536,0.04760266674889458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,1024,0.023372444841596816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,10240,0.021321778496106465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,64,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,2048,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,1536,0.0039057777159743835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,16384,0.027292443646325007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,4096,0.011855111353927188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,256,0.00527022240890397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,128,0.059564444753858775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,8192,0.2175920009613037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,256,0.0024951110697454875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,7168,0.0136808885468377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,3584,0.010402667025725046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,3072,0.020072888996866014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,2560,0.036010666026009455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,3584,0.00793599999613232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,3072,0.01000088867213991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,6144,0.04650222261746725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,256,0.06544799937142266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,256,0.014145778285132514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,512,0.00527022240890397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,16384,0.03256355391608344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,1536,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,3072,0.0038986665507157645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,12288,0.01721333298418257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,6144,0.011728888584507836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,2560,0.008177777959240807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,5120,0.009658666948477427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,128,0.001830222292078866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,2048,0.007920000288221572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,128,0.004925333377387789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,32,0.014411555396185981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,10240,0.09116088681750828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,2048,0.004596444467703502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,1536,0.016864000095261466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,2048,0.03566044569015503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,4096,0.007613333563009898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,512,0.005632888939645555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,16384,0.10614577929178874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,65536,0.16992178228166369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,16384,0.011187555889288584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,65536,0.4641760190327962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,8192,0.0110880004035102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,256,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,768,0.016159110599093966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,12288,0.039883553981781006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,12288,0.01755288905567593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,16384,0.0427706672085656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,12288,0.0181697772608863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,256,0.05972177452511258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,512,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,3584,0.011092444260915121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,3072,0.016493333710564505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,65536,0.07442577679951985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,32,0.00868622213602066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,1024,0.006991110742092133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,8192,0.025381333298153345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,6144,0.00926755534278022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,5120,0.06113955709669325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,65536,0.0790133343802558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,8192,0.010752889017264048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,2560,0.06739466720157199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,1536,0.011965333587593503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,10240,0.12105154991149902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,3072,0.013432888521088494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,256,0.007714666426181793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,1536,0.018587556150224473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,7168,0.022856000396940444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,512,0.0025173332542181015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,128,0.0151573336786694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,512,0.010650667051474253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,7168,0.013599110974205865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,6144,0.009130666653315226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,128,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,7168,0.04026755690574646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,128,0.0025315555847353404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,10240,0.052947554323408336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,2048,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,768,0.00254488881263468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,7168,0.022014222211307947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,16384,0.04864889052179125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,8192,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,7168,0.01311466677321328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,2560,0.0148044443792767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,12288,0.41031821568806964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,65536,0.7636017799377441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,4096,0.006603555546866522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,3072,0.07296978102789985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,10240,0.15229511260986328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,16384,0.043064888980653554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,7168,0.01793599956565433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,2560,0.015127999915017022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,6144,0.009011555876996782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,1536,0.00885333369175593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,32,0.0017955555684036678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,512,0.013072888884279462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,4096,0.015086222026083203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,8192,0.013343999783198038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,1024,0.011747555600272285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,65536,0.09700622161229451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,7168,0.09523466560575698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,16384,0.017297777864668105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,1024,0.060159113672044545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,5120,0.0090515555606948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,5120,0.023981332778930664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,8192,0.011049778097205691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,7168,0.012668444050682915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,32,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,1024,0.016534222496880423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,4096,0.020495999190542433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,65536,0.7838017675611707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,1536,0.005929777605666055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,4096,0.030801776382658217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,256,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,2048,0.006674666785531574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,7168,0.010077333284748925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,4096,0.006583111153708563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,16384,0.048727999130884804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,768,0.00544888898730278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,2048,0.004593777573770947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,256,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,1024,0.022648889157507155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,128,0.027003554834259882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,7168,0.46088621351453996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,4096,0.023539554741647508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,5120,0.011317333413494958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,1024,0.025166221790843542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,512,0.05423911081420051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,10240,0.006235555642180973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,128,0.010425777898894416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,3584,0.015206222732861837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,7168,0.013109332985348172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,6144,0.014365333649847241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,256,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,8192,0.08224088615841336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,3584,0.007630222373538547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,16384,0.0324924455748664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,768,0.041478223270840116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,2048,0.04343733191490173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,1536,0.022712000542216833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,32,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,16384,0.051038222180472485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,128,0.011332444846630096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,4096,0.01180355581972334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,512,0.004895110925038655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,6144,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,32,0.009728888670603434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,1024,0.004967111266321606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,512,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,3584,0.00794044468137953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,6144,0.011602666642930774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,256,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,32,0.006559111177921295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,3584,0.007420444654093848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,8192,0.015104000767072042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,10240,0.010110222631030613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,4096,0.007913777397738563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,1536,0.005892444401979446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,768,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,1536,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,768,0.006272000157170826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,16384,0.04691466689109802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,768,0.0038444445365005066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,12288,0.013787555197874704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,5120,0.019726221760114033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,6144,0.013510222236315409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,32,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,1536,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,3584,0.015150222513410779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,64,0.010341333018408881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,4096,0.011069333387745751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,12288,0.05992088715235392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,7168,0.029270221789677937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,3584,0.015442666080262927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,2560,0.01761866609255473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,3072,0.007335111498832703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,12288,0.010185778141021729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,7168,0.029595557186338637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,16384,0.027797334724002417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,10240,0.021708443760871887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,3584,0.018233777748213876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,16384,0.1361617777082655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,1024,0.013819555441538492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,3584,0.02937244375546773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,128,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,2560,0.010220444036854638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,2560,0.006657777975002925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,32,0.005583999885453119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,6144,0.012078222301271228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,6144,0.1295751068327162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,6144,0.03288711110750834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,2560,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,16384,0.10638488663567437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,1024,0.005865777946180767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,3072,0.010524444282054901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,8192,0.02608000073168013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,2560,0.019828443725903828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,1536,0.004027555386225383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,768,0.006184000107977126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,5120,0.01275466630856196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,128,0.0027964444210131965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,2048,0.013769778112570444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,16384,0.04377777708901299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,1536,0.015605333778593274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,32,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,768,0.005962666538026597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,5120,0.016513778103722464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,16384,0.03319999906751845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,65536,0.1230506632063124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,8192,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,256,0.006213333457708359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,512,0.006258666515350342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,6144,0.02536800007025401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,7168,0.012950222525331708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,3072,0.006917333437336816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,16384,0.04888977938228183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,1536,0.003536000019974179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,3072,0.11800177892049153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,1024,0.008640888664457533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,3072,0.01702222228050232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,6144,0.025655999779701233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,256,0.016157334049542744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,7168,0.08579822381337483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,5120,0.015977778368526034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,2048,0.005280888742870755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,3584,0.03526399864090814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,768,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,16384,0.06494488981035021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,8192,0.014655111564530266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,65536,0.23144978947109648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,2560,0.006594666590293248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,1536,0.0038862224254343244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,768,0.03326133224699233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,10240,0.05613155497445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,65536,0.06429244412316217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,4096,0.007602666815121968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,512,0.0029608888758553397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,12288,0.004877333425813251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,6144,0.019140443868107267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,10240,0.026955554882685345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,3584,0.007556444240940942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,16384,0.08879288699891831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,6144,0.041998220814598926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,65536,0.024509333901935156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,16384,0.01959466603067186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,768,0.020603555772039626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,3584,0.008048888709810045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,3584,0.009371555513805812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,128,0.0021137777302000257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,64,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,3072,0.009920888476901585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,5120,0.00867555538813273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,2048,0.028763555818133887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,512,0.003559999995761447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,10240,0.08477866649627686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,7168,0.27154845661587185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,3584,0.009249777429633671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,4096,0.0391902228196462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,16384,0.05245688888761732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,1536,0.003871110992299186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,128,0.004857777721352047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,1536,0.01787644459141625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,1024,0.027525333894623652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,10240,0.029842668109469946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,8192,0.02010222276051839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,512,0.01514488955338796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,4096,0.014803555276658801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,2560,0.006966222491529252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,6144,0.03306488858328925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,6144,0.009657777845859528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,3072,0.010528888967302112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,7168,0.021665778425004747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,2560,0.004913777940803104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,64,0.006228444476922353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,64,0.0021262222694026097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,3584,0.01479200025399526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,768,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,512,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,8192,0.3446186648474799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,64,0.002143111079931259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,10240,0.016542222764756944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,16384,0.020502222908867728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,6144,0.02233155568440755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,768,0.023012444376945496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,8192,0.03108888864517212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,3584,0.008443555898136562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,768,0.005330666485759947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,8192,0.013536888692114087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,4096,0.016158221496476066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,256,0.014761777387724983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,16384,0.08288177516725329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,3584,0.014256000518798828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,768,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,512,0.09270044167836507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,768,0.009855111440022787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,2560,0.010715555813577441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,8192,0.05043111244837443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,3072,0.011929777761300405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,4096,0.004295111116435793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,2048,0.004550222307443619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,12288,0.0336133340994517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,4096,0.010919111470381418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,128,0.014473777678277759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,12288,0.1633111106024848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,3584,0.00794311116139094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,64,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,5120,0.009658666948477427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,2048,0.007974222302436829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,256,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,1536,0.007736889024575551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,10240,0.02574133376280467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,1536,0.018586667047606576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,128,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,2048,0.006846222198671765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,3072,0.006973333656787872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,1024,0.005120000077618493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,3584,0.010060444474220276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,256,0.014155555102560254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,4096,0.02022133270899455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,64,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,3584,0.021659556362364028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,5120,0.008301333420806462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,1536,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,32,0.01033688916100396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,3072,0.019381332728597853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,128,0.013527111046844058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,1024,0.004870222260554631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,8192,0.009300444689061906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,256,0.01312622262371911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,1536,0.006976889239417181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,3072,0.018813333577579923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,10240,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,768,0.005094222310516569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,2048,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,2560,0.04792266752984789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,256,0.002520888836847411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,64,0.006262222097979651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,8192,1.5521920522054036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,16384,0.005535999933878581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,5120,0.016013332539134555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,7168,0.02597155504756504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,4096,0.01573422220018175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,1536,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,2048,0.0042204442951414315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,128,0.0021102221475707158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,768,0.017870222528775532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,7168,0.06849244568083021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,6144,0.02647200061215295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,10240,0.9307528601752387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,10240,0.016207110550668504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,256,0.03302133414480422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,1536,0.012763555679056378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,16384,0.9046452840169271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,5120,0.02314044369591607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,7168,0.038322667280832924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,2560,0.0240328891409768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,5120,0.025840888420740765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,3072,0.018750222192870248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,512,0.023919110496838886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,128,0.022092445029152766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,2048,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,12288,0.011597333682907952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,1536,0.006258666515350342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,16384,0.09506044122907852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,512,0.01187111106183794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,256,0.0539608895778656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,10240,0.02160355614291297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,6144,0.00904622260067198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,12288,0.023185777995321486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,12288,0.012773333324326409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,2560,0.021377777059872944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,1536,0.05298044284184774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,768,0.01481066644191742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,6144,0.013440888788965015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,5120,0.026566222310066223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,128,0.0019519999623298645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,2048,0.019533332851197984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,6144,0.00906755526860555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,7168,0.02628622286849552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,7168,0.034727109803093806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,3584,0.09477955765194362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,3072,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,10240,0.011786667009194693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,2048,0.007264888948864407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,1536,0.024877332978778418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,6144,0.008822222550710043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,16384,0.03536977701716953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,512,0.018767111831241183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,256,0.00488800017370118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,2560,0.051536887884140015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,128,0.009999111294746399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,1024,0.0031644445326593188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,8192,0.025760889053344727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,4096,0.008389333056079017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,4096,0.011503111157152386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,3584,0.007720000214046902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,1536,0.005389333185222413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,512,0.020937777227825586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,2048,0.006250666661394968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,768,0.00658044425977601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,512,0.012648888760142855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,2560,0.00830577810605367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,1024,0.01698755555682712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,5120,0.04052977760632833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,256,0.011790222591824003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,1024,0.004384888956944148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,512,0.005542222410440445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,16384,0.010013333625263637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,3584,0.043952888912624784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,512,0.016203555795881484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,64,0.006252444452709622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,3584,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,3584,0.02367466688156128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,64,0.0019333333604865605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,65536,0.2840622266133626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,3584,0.007615110940403408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,768,0.005397333453098933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,8192,0.05664177735646566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,16384,0.025051555699772302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,2560,0.01352533366945055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,8192,0.04192355606291029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,128,0.00703111125363244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,256,0.011696000066068439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,1024,0.005756444401211209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,128,0.005919999960396025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,10240,0.01758399936887953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,2560,0.006798222247097228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,1536,0.01310488912794325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,1024,0.03296711047490438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,6144,0.01402311192618476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,128,0.014133334159851074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,10240,0.22455644607543945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,2560,0.022064000368118286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,2560,0.006587555425034628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,8192,0.012404444317022959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,3072,0.020079111059506733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,3072,0.025031111306614343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,6144,0.023380445109473333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,2560,0.016484444340070088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,8192,0.035622222556008235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,4096,0.04087466663784451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,6144,0.07194133599599202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,768,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,4096,0.007374222079912822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,4096,0.007650666766696506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,2560,0.007309333317809635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,10240,0.012057777908113269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,768,0.006211555666393704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,768,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,8192,0.03538577755292257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,512,0.005391110976537068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,3584,0.0240639994541804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,1536,0.0127697777416971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,5120,0.0366977784368727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,64,0.0018515555808941524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,2048,0.01921688848071628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,512,0.0039013334446483185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,65536,0.05675377448399862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,6144,0.01572355628013611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,8192,0.028981334633297388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,512,0.00527555536892679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,128,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,1024,0.0029973333908451926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,1024,0.005807999935415056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,256,0.014476444986131458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,7168,0.1352284484439426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,4096,0.007664889097213745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,128,0.1944204436408149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,4096,0.015272888872358533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,4096,0.008750222623348236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,512,0.0035208890007601846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,128,0.014179555906189812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,768,0.0262746661901474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,16384,0.02363555630048116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,4096,0.006979555719428592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,512,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,3584,0.004943111290534337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,16384,0.017312889297803242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,1024,0.0038986665507157645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,2560,0.025434666209750708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,256,0.018437332577175565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,3072,0.010954666468832227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,7168,0.01237155579858356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,4096,0.00793599999613232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,4096,0.007625777688291337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,5120,0.07845511039098103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,64,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,65536,0.08288088772031996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,2048,0.019137778215938143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,7168,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,256,0.009017777939637503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,2048,0.023965333898862202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,12288,0.030416889323128596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,5120,0.008632000121805403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,16384,0.051426665650473706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,3584,0.011716444459226398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,1536,0.017840888765123155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,2560,0.010112889111042023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,65536,0.06880444288253784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,2048,0.08545777532789443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,2048,0.006223111102978389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,3584,0.007708444363541073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,768,0.002965333354141977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,4096,0.013803555733627744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,768,0.007072000039948358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,3072,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,128,0.015152000718646579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,3072,0.019555555449591745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,16384,0.07420977618959215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,10240,0.27314578162299263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,768,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,7168,0.011731555892361535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,768,0.005632888939645555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,3072,0.009691555466916826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,256,0.026404443714353774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,10240,0.033879111210505165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,10240,0.02087822225358751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,65536,5.308843400743273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,12288,0.014904000692897372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,32,0.029471112622155085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,10240,0.0565519995159573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,512,0.0028880000528362063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,3072,0.017839110559887357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,8192,0.042797333664364286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,10240,0.03526844580968221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,1024,0.007991111112965478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,512,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,1536,0.0038488888078265717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,7168,0.01993600030740102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,768,0.0028479999552170434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,3072,0.007260444263617198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,4096,0.010565333068370819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,12288,0.267922666337755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,256,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,1536,0.009916444619496664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,6144,0.021240888370407954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,6144,0.009888889061080085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,3072,0.016892444756295946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,3072,0.009020444419648912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,768,0.007224889265166388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,256,0.02643111182583703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,1536,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,16384,0.030652443567911785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,10240,0.05146400133768717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,64,0.009669333696365356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,1536,0.022834666901164587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,3584,0.009155555731720394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,65536,0.044575111733542554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,1536,0.03958933220969306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,5120,0.008619555168681675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,16384,0.04111466805140177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,1536,0.006579555571079254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,1536,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,12288,0.014108444253603617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,512,0.003552000141806073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,7168,0.028499556912316218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,2560,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,2560,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,2560,0.027423111928833857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,256,0.020080000162124634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,2048,0.12763110796610513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,32,0.005912000106440649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,128,0.002191999927163124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,16384,0.38025689125061035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,16384,0.042903999487559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,8192,0.01072533345884747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,5120,0.25015200508965385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,1024,0.006590222318967183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,10240,0.011755555868148804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,1536,0.01889866590499878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,6144,0.10564711358812119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,3072,0.009992000129487779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,5120,0.010703110860453712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,12288,0.01549600064754486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,128,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,512,0.003165333428316646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,65536,0.05611199802822537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,128,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,2048,0.00655377780397733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,1536,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,8192,0.010348444183667501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,5120,0.01792711185084449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,1024,0.08326933119032118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,8192,0.017235555582576327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,2048,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,16384,0.005895111295912001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,7168,0.016326222154829238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,7168,0.11650310622321235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,1536,0.006351111249791251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,128,0.014512000812424554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,256,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,256,0.005209777918126848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,32,0.0018497777895794974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,2560,0.009387555221716562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,128,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,512,0.003112888791494899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,64,0.002485333424475458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,65536,0.08053777615229289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,3072,0.013724444641007317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,5120,0.07433778047561646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,64,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,256,0.02439111140039232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,512,0.016604445046848722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,1536,0.00794577764140235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,128,0.055735111236572266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,128,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,7168,0.093341330687205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,6144,0.009178666604889764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,6144,0.018182223041852314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,12288,0.02052177819940779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,1024,0.003528888854715559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,32,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,6144,0.030568000343110826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,768,0.01874755488501655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,64,0.002809777855873108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,4096,0.010746666954623329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,2560,0.026849778162108526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,2048,0.006298666612969504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,768,0.015141333142916361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,128,0.010001777774757808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,128,0.01275466630856196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,2048,0.007144889069928064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,256,0.004187555362780889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,8192,0.009326222042242685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,256,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,128,0.010048000348938836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,1024,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,1536,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,3072,0.00814666681819492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,512,0.004634666774008009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,7168,0.012378666963842181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,10240,0.0499857763449351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,3584,0.011730666789743634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,4096,0.007984889050324758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,3584,0.01107022249036365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,4096,0.010527110762066312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,5120,0.010503999888896942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,2560,0.006646222124497096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,65536,0.14149867163764104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,512,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,12288,0.01612711118327247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,4096,0.015568888849682279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,2560,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,2048,0.02530488868554433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,512,0.003169777699642711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,2048,0.013435555828942193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,2560,0.020581333173645865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,5120,0.01975022256374359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,12288,0.03546844588385688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,12288,0.051968889103995435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,16384,0.027127110295825537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,12288,0.044565333260430225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,3584,0.009098666409651438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,2048,0.007269333634111617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,5120,0.02424266603257921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,512,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,12288,0.013911111487282647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,65536,0.08634133471382989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,1024,0.007650666766696506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,5120,0.013289777768982781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,2560,0.031239999665154353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,1024,0.021607110897699993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,512,0.005294222384691238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,16384,5.2032695346408415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,16384,0.01977155605951945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,1024,0.018231999542978074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,1024,0.01514933341079288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,7168,0.09577333264880711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,2560,0.006608888920810487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,3072,0.00416711096962293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,2048,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,8192,2.6539279090033636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,7168,0.0274000002278222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,7168,0.0717031094763014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,4096,0.03859733210669623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,2560,0.009029332962301042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,10240,0.008423999779754216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,8192,0.010678222609890832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,1536,0.005954666684071223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,10240,0.02919644448492262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,12288,0.013395555317401886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,512,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,16384,0.029845333761639063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,256,0.004933333231343163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,5120,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,3072,0.007056000332037608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,3072,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,3584,0.023132445083724126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,12288,0.021440888444582622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,12288,0.058296885755327016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,6144,0.04013333386845059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,1024,0.005663110978073544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,128,0.004923555586073133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,512,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,12288,0.04232177800602383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,65536,0.01586933268441094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,65536,0.22240000300937227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,1536,0.003556444413132138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,2048,0.05161777800983853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,12288,0.3478071159786648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,256,0.024638222323523626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,3072,0.00700444479783376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,4096,0.00793422261873881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,65536,0.18331378036075166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,3072,0.025598223010698955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,7168,0.015420445137553744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,1024,0.012610666453838348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,1024,0.00553688903649648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,3072,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,512,0.0034933334423436057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,256,0.016487111647923786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,128,0.006024000131421619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,128,0.00488800017370118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,6144,0.017482666505707633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,3584,0.007638221813572778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,3072,0.007601777712504069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,65536,0.20260266462961832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,12288,0.015850666496488783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,8192,0.020242666204770405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,4096,0.03394488824738397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,256,0.007358222372002072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,65536,0.059776001506381564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,8192,0.023680888944202002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,3072,0.009709333380063375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,8192,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,32,0.005587555468082428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,768,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,128,0.005252444495757421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,2048,0.049504889382256397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,8192,0.015357333752844067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,128,0.014490667316648694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,64,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,3584,0.02125688890616099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,1536,0.019894222418467205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,512,0.00488800017370118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,3584,0.03049066662788391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,128,0.01447111037042406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,3072,0.0052151112920708126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,10240,0.020908445119857788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,2048,0.09971821970409817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,1024,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,512,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,512,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,3072,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,65536,0.09771911303202312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,10240,0.046612444851133555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,1536,0.01614488826857673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,512,0.004870222260554631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,256,0.0063973334100511335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,7168,0.0707919994990031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,1536,0.004571555389298333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,3072,0.017804443836212158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,5120,0.008618666893906063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,128,0.005962666538026597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,65536,0.09690044323603313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,10240,0.05115911033418444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,3584,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,8192,0.028629332780838013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,6144,0.009709333380063375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,256,0.005028444445795483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,3584,0.011055111057228513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,256,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,5120,0.03621066610018412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,128,0.002135111019015312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,1536,0.013053333593739403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,2048,0.014536889062987434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,2560,0.009011555876996782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,8192,0.018544889158672757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,1024,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,128,0.013759999639458127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,3072,0.024825778272416856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,64,0.005881777654091517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,1024,0.015456888410780164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,5120,0.009000889129108852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,5120,0.11152088642120361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,12288,0.04313955704371134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,768,0.003185777821474605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,1536,0.018174222773975797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,32,0.00425511101881663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,8192,0.011712888876597086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,16384,0.033265779415766396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,1024,0.018872888551818002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,8192,0.004949333353175057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,7168,0.0347671111424764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,3072,0.007330666813585494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,65536,0.0703164471520318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,2560,0.006592000110281839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,12288,0.009992889232105678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,2048,0.014814222852389017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,12288,0.7584035661485461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,2560,0.006909333169460297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,768,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,7168,0.03293155630429586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,4096,0.007663110891977946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,1024,0.017258667283587985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,32,0.004923555586073133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,2048,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,6144,0.00620888877246115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,5120,0.01851733360025618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,16384,0.04266311062706841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,512,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,128,0.0021359999146726397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,3584,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,6144,0.009104889300134447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,12288,0.039135999149746366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,4096,0.008632889224423302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,1024,0.005879111174080107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,32,0.0021199999998013177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,768,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,2048,0.006232888748248418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,2560,0.007292444507280986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,6144,0.1730008920033773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,3072,0.02684177789423201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,2560,0.0248391115003162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,256,0.009350222018029954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,4096,0.03497955534193251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,10240,0.6593093342251247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,512,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,3072,0.019201777047581144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,3072,0.016563556260532804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,3072,0.007360888852013483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,6144,0.13162133428785536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,3584,0.009834667046864828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,768,0.016922665966881644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,10240,0.030638221237394545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,4096,0.015765332513385348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,4096,0.07372977998521593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,1024,0.03201155530081855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,8192,0.020080889264742535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,5120,0.009519110951158736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,64,0.001842666624320878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,4096,0.016223111086421542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,256,0.004888888862397936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,512,0.021426666114065383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,128,0.004567111117972268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,32,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,3072,0.027523555689387854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,16384,0.022985777921146814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,6144,0.004603555632962121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,768,0.005583111196756363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,2560,0.007263110743628607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,3072,0.017285333739386666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,6144,0.02755911151568095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,1536,0.01108977778090371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,6144,0.017172444197866652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,10240,0.023077333966890972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,10240,0.050372444921069674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,1536,0.00811288919713762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,768,0.004521777646409141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,8192,0.010689777632554373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,65536,0.13078667057885066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,65536,0.5299919976128472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,768,0.00564177789621883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,65536,0.11575555801391602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,5120,0.016184889607959323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,5120,0.21307022041744658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,1024,0.026476444469557867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,512,0.005251555393139522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,2048,0.02311822275320689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,512,0.009737778041097853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,4096,0.007872888611422645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,5120,0.008294222255547842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,2560,0.008112000094519721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,128,0.004549333204825719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,12288,0.027271999253167048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,768,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,4096,0.10081155432595147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,512,0.003440888805521859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,512,0.01584533353646596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,3072,0.023456000619464453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,768,0.012728888955381183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,2048,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,1024,0.006532444308201472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,128,0.009336888790130615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,8192,0.01498044364982181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,32,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,12288,0.01670488880740272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,256,0.002171555534005165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,512,0.017789334058761597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,2560,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,10240,0.020231111182106864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,65536,0.015795555379655626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,3072,0.008833777573373582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,3072,0.01477599971824222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,2560,0.007063111497296228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,128,0.005664000080691443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,16384,0.027818666564093694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,3072,0.012698666916953193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,10240,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,128,0.005336888962321811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,3584,0.007265778051482306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,16384,0.016881777180565726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,5120,0.10600622495015462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,10240,0.031108445591396753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,7168,0.01921066641807556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,6144,0.02070577773782942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,3584,0.010017777482668558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,64,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,10240,0.04510666595564949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,3584,0.006611555400821898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,16384,0.481569766998291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,2048,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,7168,0.014601776997248331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,1024,0.006959999601046245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,768,0.009470221896966299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,8192,0.029458665185504492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,512,0.003953777667548922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,768,0.007956444389290279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,8192,0.021349334054523047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,768,0.0028497777465316984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,1024,0.006020444548792309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,6144,0.0543120006720225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,256,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,3584,0.007607999775144789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,2560,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,64,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,65536,0.07553599940405951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,1024,0.00591111100382275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,32,0.0018240000224775737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,12288,0.018936889039145578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,3072,0.007244444555706448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,12288,0.021202666891945735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,4096,0.009441778063774109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,4096,0.018194667167133756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,5120,0.01463822192615933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,12288,0.04474844535191854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,4096,0.018778666853904724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,4096,0.011553777588738335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,16384,0.039020445611741804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,512,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,512,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,1536,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,32,0.0018195555441909367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,8192,0.01367111090156767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,5120,0.010693333215183683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,128,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,512,0.002483555633160803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,256,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,3584,0.010211555494202508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,7168,0.022052443689770166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,64,0.0018568889548381169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,512,0.002529777793420686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,5120,0.031074666314654883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,3584,0.009780444204807281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,3072,0.010693333215183683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,65536,0.152456005414327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,6144,0.008638222184446123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,65536,0.25523821512858075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,1536,0.018248000078731112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,16384,0.1813911067114936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,1024,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,256,0.0028924445311228433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,1024,0.16185511483086482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,1024,0.0188746667570538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,4096,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,6144,0.007191110816266801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,10240,0.03286400106218126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,128,0.005537777725193236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,7168,0.11592266294691299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,2048,0.01889600025282966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,65536,0.19694133599599203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,6144,0.010378667049937779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,2048,0.38719643486870664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,65536,0.006278222219811545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,16384,0.06597333484225802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,1024,0.013752889302041797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,12288,0.004859555512666702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,7168,0.009069333473841349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,768,0.007276444799370236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,512,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,6144,0.1289342244466146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,5120,0.016211556063758004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,65536,0.07879733377032809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,1024,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,512,0.002471111093958219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,3072,0.0884115563498603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,1024,0.0058559998869895935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,1024,0.006632888896597757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,1024,0.015832000308566622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,6144,0.03495288888613383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,8192,0.015182221929232279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,4096,0.007939555578761632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,2048,0.022279111875428095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,768,0.014838222000333997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,4096,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,3584,0.005050666630268097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,128,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,4096,0.01586844523747762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,128,0.002161777681774563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,7168,0.013428444663683573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,4096,0.11786844995286729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,7168,0.021026665965716045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,128,0.004870222260554631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,65536,0.06951822174919976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,2560,0.006340444501903322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,768,0.0057706667317284485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,2560,0.025449777642885845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,6144,0.26144178708394367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,3584,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,3072,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,3072,0.007328888608349695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,65536,1.0958195792304146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,2048,0.05786489115820991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,6144,0.047042667865753174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,5120,0.031239999665154353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,4096,0.02862488892343309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,1024,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,7168,0.05520977907710605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,8192,0.19600178135765922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,512,0.01277955538696713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,32,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,5120,0.029309332370758057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,1024,0.011711111499203576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,10240,0.4343840016259088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,5120,0.01518133282661438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,10240,0.024671110841963027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,10240,0.021377777059872944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,16384,0.04350399971008301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,768,0.031135112047195435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,1024,0.004217777815130022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,12288,0.06443466742833455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,2560,0.01479200025399526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,2048,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,512,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,16384,0.04093955622778998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,1024,0.012103111379676394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,6144,0.012729778057999082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,2048,0.01366311146153344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,16384,0.09887466828028361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,10240,0.009415999882751042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,2560,0.03330844309594896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,1024,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,10240,0.06906577613618639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,1024,0.018589332699775696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,4096,0.00796088907453749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,128,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,2560,0.006913777854707506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,4096,0.025719111164410908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,16384,0.16691554917229545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,10240,0.03190666768285964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,3584,0.008447111480765873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,2048,0.0042275554604000514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,6144,0.00656355544924736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,512,0.003528888854715559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,512,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,5120,0.012070222033394707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,768,0.012726222475369772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,128,0.014785778191354541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,64,0.0018462222069501877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,256,0.0028942223224374983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,64,0.07760800255669488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,8192,0.014828445182906257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,32,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,6144,0.16743821567959258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,4096,0.008962666822804345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,10240,0.09828444321950276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,12288,0.04800800151295132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,4096,0.02421955598725213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,32,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,1024,0.005879111174080107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,16384,0.02404888967672984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,16384,0.026305778159035578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,10240,0.010661332971519895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,65536,0.088191999329461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,3584,0.20394133196936715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,32,0.007258666886223688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,1024,0.009639999932712978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,2560,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,10240,0.09617333279715644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,7168,0.005421333428886201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,4096,0.007650666766696506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,5120,0.00869511150651508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,6144,0.07717955774731107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,256,0.006255110932721033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,4096,0.06057422028647529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,8192,0.0192275560564465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,12288,0.016063110695944894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,4096,0.01756000022093455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,6144,0.010223111344708337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,1536,0.003863999827040566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,10240,0.01313511116637124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,4096,0.016826666063732572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,65536,0.1193102200826009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,1536,0.08154488934410943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,3584,0.16417955027686224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,7168,0.014412444498803882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,7168,0.036350223753187395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,3584,0.007607111500369177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,1536,0.16919999652438691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,512,0.016583111551072862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,768,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,2048,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,16384,0.029863999949561224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,12288,0.041209777196248375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,65536,0.07204533285564847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,16384,0.10969777901967366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,8192,0.015504888362354703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,16384,0.2995813422732883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,12288,0.017906667457686532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,32,0.005384888913896348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,65536,0.09605333540174697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,2560,0.025410667061805725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,3584,0.0981724460919698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,2048,0.056208001242743603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,5120,0.03493155704604255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,3072,0.015626667274369132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,32,0.0018159999615616268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,768,0.006234666539563074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,7168,2.604649861653646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,256,0.0028195555011431375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,256,0.01824711097611321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,512,0.014841778410805596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,4096,0.026313778426912095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,12288,0.058392001522911914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,5120,0.03587288988961114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,7168,0.0326275560590956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,1024,0.020900444851981267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,10240,0.010496888723638324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,65536,0.07382844554053412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,16384,0.043881777260038585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,16384,0.018927110566033255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,2048,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,512,0.014497778481907315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,6144,0.01909244391653273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,3584,0.004930666751331753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,768,0.03307911091380649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,512,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,7168,0.023455111516846552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,12288,0.02127200033929613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,64,0.0028275555620590844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,2048,0.03682311044798957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,256,0.01588711142539978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,10240,0.04981599913703071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,65536,0.09060533179177178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,1024,0.005543111099137201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,768,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,256,0.004895110925038655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,512,0.010677333507272931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,3072,0.0041777777175108595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,4096,0.008276444342401292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,256,0.05577244361241659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,3072,0.007272889216740926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,3584,0.00794933322403166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,10240,0.012061333490742577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,768,0.016474665866957772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,6144,0.009095999929640029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,5120,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,512,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,128,0.1044515570004781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,4096,0.011868444581826528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,8192,0.02586666742960612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,2048,0.004222222086456087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,768,0.004928888960017098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,10240,0.012087999946541257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,8192,0.10104711188210382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,12288,0.01313777764638265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,1536,0.00797155582242542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,3072,0.04818311002519396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,2048,0.006619555668698416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,3072,0.013668444421556262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,256,0.037422223223580256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,128,0.007987555530336168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,768,0.0032071111102898917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,2560,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,1024,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,64,0.007678222325113084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,12288,0.019779556327395968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,8192,0.014656888114081489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,6144,0.07006488906012641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,32,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,128,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,1024,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,512,0.012431999875439538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,1024,0.003181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,4096,0.0352168877919515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,3072,0.010114666488435533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,16384,0.01718755563100179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,65536,0.10303378105163574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,16384,1.4057910707261827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,512,0.030712889300452337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,4096,0.008244444098737504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,1024,0.006024000131421619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,32,0.005589333259397083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,3584,0.007302222152551015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,65536,0.07883110973570082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,16384,0.14686044057210287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,8192,0.0123857781291008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,3072,0.007257777783605788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,3072,0.007296889192528195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,768,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,512,0.005656888915432825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,3584,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,1024,0.011698666546079846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,1536,0.006285333385070165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,256,0.005953777581453323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,768,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,3072,0.026760889424218073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,4096,0.02518133322397868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,3584,0.039135999149746366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,5120,0.008644444247086843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,2560,0.08206044303046332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,16384,0.016397333807415433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,12288,0.02756355537308587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,768,0.008314666648705801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,128,0.004104888687531154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,768,0.003506666670242945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,6144,0.03379644287957086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,2560,0.025613332788149517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,4096,0.01568800045384301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,512,0.008992888861232335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,3072,0.010006222460005019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,65536,0.12527644634246826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,5120,0.015220445063379077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,128,0.00499288903342353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,8192,0.01647555496957567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,7168,0.04116800096299913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,10240,0.06414666440751818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,1536,0.01142133358452055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,12288,0.048914667632844716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,2048,0.006529777828190062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,3072,0.007659555309348636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,12288,0.02472977836926778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,6144,0.14916888872782388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,4096,0.008171555896600088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,12288,0.014115555418862237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,7168,0.04825599988301595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,5120,0.011241777903503843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,65536,1.4737075169881184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,2560,0.02330844435426924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,512,0.003528000166018804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,1536,0.007982222570313348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,512,0.002524444419476721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,7168,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,256,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,65536,0.08139110936058892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,256,0.008002666963471307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,5120,0.00888888869020674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,64,0.001834666563404931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,1024,0.0028906667398081887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,1536,0.05383200115627713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,1024,0.006631999793979857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,4096,0.03141155507829454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,7168,0.009907555249002244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,65536,0.03576622075504727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,8192,0.06003377834955851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,8192,0.12148622671763103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,8192,0.0127697777416971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,2560,0.007240888973077138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,2560,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,256,0.007620444728268518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,5120,0.008840888738632202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,128,0.014111111561457316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,8192,0.014255111416180929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,3072,0.024896888269318476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,3584,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,16384,0.020510221521059673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,7168,0.07594044340981378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,1024,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,768,0.0028026666906144884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,16384,0.03299466768900553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,8192,0.01089866707722346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,1536,0.006976889239417181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,3584,0.07039911217159696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,1024,0.0177831103404363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,512,0.006229333165619109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,16384,0.21838043795691597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,3072,0.009319999979601966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,64,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,6144,0.020992888344658747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,256,0.016234666109085083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,1536,0.018030222919252183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,3584,0.013830222189426422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,7168,0.018539556198649935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,2048,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,4096,0.009331555830107795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,256,0.013796444568369122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,7168,0.020583111378881667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,5120,0.008465777668688033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,7168,0.03503911031617059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,3584,0.007311999797821045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,3072,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,3072,0.011215111447705163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,3072,0.017439999514155917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,256,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,10240,0.01678311162524753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,2560,0.04903911219702827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,256,0.004713777866628435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,7168,0.018581334087583754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,1024,0.0028684443483750024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,1536,0.010308444499969482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,16384,0.041434665520985924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,16384,0.026895999908447266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,6144,0.14289511574639216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,6144,0.0090524446633127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,2560,0.022653333014912073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,7168,0.029065777858098347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,12288,0.0692853331565857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,3072,0.014112000664075216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,10240,0.011696889168686338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,2048,0.02864266766442193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,768,0.018907555275493197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,768,0.014142221874660917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,1024,0.015817777978049386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,4096,0.09070577886369492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,128,0.01238844460911221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,4096,0.020246222615242004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,16384,0.6266604529486762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,10240,0.11019111341900295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,16384,0.26249776946173775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,4096,0.010784889260927836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,10240,0.0510844455824958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,7168,0.014766222900814481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,2048,0.006218666831652324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,3584,0.015777778294351365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,3072,0.014081777797804939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,3072,0.014839111102951897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,512,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,2560,0.019653333557976615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,6144,0.11450400617387559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,6144,0.009641778137948778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,128,0.0031804444475306403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,5120,0.012162666353914471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,12288,0.006295111030340195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,1024,0.006228444476922353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,768,0.011750222080283694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,2048,0.006268444574541516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,65536,0.277200010087755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,3584,0.01070844464831882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,8192,0.02533422244919671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,12288,0.013728888498412238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,128,0.009986666341622671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,512,0.015856888559129503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,16384,0.33486045731438535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,5120,0.010028444230556488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,6144,0.017282666431532968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,32,0.002141333288616604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,10240,0.019201777047581144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,2048,0.017818666166729398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,3584,0.015989333391189575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,1024,0.0038568890757030914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,12288,0.035644445154401995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,6144,0.30727465947469074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,32,0.004902222090297275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,65536,0.18201422691345215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,2048,0.0053173332578606075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,10240,0.04352444410324097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,7168,0.009963555468453301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,1536,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,768,0.017480888300471835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,5120,0.01681866745154063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,256,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,7168,0.022637334134843614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,8192,0.011757333245542316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,3072,0.013064888616402945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,8192,0.09947733084360759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,4096,0.012714666624863943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,1024,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,2048,0.05519466598828634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,2048,0.012858666479587555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,2048,0.00832799987660514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,10240,0.036363555325402155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,5120,0.020079111059506733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,8192,0.02160088883505927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,512,0.018182223041852314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,4096,0.239574220445421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,1536,0.03404533200793796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,16384,0.07524533404244317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,5120,0.10862577623791164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,6144,0.009015999734401703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,4096,0.023002665903833177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,6144,0.15524178081088597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,512,0.01717866626050737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,3584,0.014375999569892883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,10240,0.013841778039932251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,4096,0.017257778180970084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,16384,0.02779910961786906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,65536,0.1608897712495592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,64,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,16384,0.030248890320460003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,10240,0.09000977542665269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,65536,0.7868000136481391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,2560,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,10240,0.028119110398822363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,512,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,2048,0.009874666730562845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,2560,0.014742222097184924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,128,0.04349422123697069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,768,0.022262222237057153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,8192,0.026176888081762526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,7168,0.15119377772013345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,65536,0.2755511071946886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,3584,0.30727020899454754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,6144,0.012725333372751871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,3584,0.015949333707491558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,4096,0.007344889144102733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,4096,0.010455999937322406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,8192,0.017338666650984023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,16384,0.020277332928445604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,256,0.016138666205936007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,3072,0.01132266637351778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,6144,0.00965244405799442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,32,0.002193777718477779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,2560,0.004929777648713854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,64,0.0018444444156355327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,32,0.006227555374304454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,16384,0.014916444818178812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,65536,0.027055111196306016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,8192,0.018935110833909776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,32,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,1536,0.003536000019974179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,5120,0.2940888934665256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,2048,0.014115555418862237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,16384,0.014448889427714877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,5120,0.02124622298611535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,7168,0.01992177797688378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,4096,0.015535111228624979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,3584,0.02898400028546651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,16384,0.018368888232443068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,8192,0.013609777722093793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,6144,0.020992888344658747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,10240,0.20131377379099527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,3584,0.007072000039948358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,65536,0.04752177662319607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,512,0.004964444372389051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,256,0.004940444396601783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,2048,0.032271110349231295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,512,0.030672887961069744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,3072,0.011050666371981302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,512,0.006280888699822956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,5120,0.02072888943884108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,6144,0.023207111491097346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,1024,0.0031999999450312722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,2048,0.007289778027269576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,65536,0.2552764415740967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,3584,0.0215013325214386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,12288,0.026735110415352717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,5120,0.01107022249036365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,16384,0.025457777910762366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,4096,0.10419022374682957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,4096,0.004961777892377642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,8192,0.0370488895310296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,65536,0.07833421892589994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,1536,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,512,0.0028862222615215513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,32,0.0032177778581778207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,512,0.02254666719171736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,4096,0.0147733340660731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,2048,0.027091556125217017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,2048,0.011701333026091257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,65536,0.8165324529012045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,4096,0.009715555442704095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,4096,0.02974133359061347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,2048,0.01854222185081906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,256,0.005952888892756567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,512,0.0025235555238193935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,2048,0.006604444649484422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,1536,0.006635555376609166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,16384,0.011392888923486074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,6144,0.011585777832402123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,16384,0.06415644619199964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,128,0.01442577772670322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,768,0.07936888933181763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,12288,0.17764888869391549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,65536,0.15721955564286974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,65536,0.24390665690104166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,16384,0.020579554968410067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,8192,0.01073066641887029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,3584,0.01962488889694214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,5120,0.027830223242441814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,5120,0.016474665866957772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,8192,0.015171556009186639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,5120,0.1294462283452352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,1536,0.018276444739765592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,8192,0.013750221994188098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,768,0.010428444378905825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,1536,0.010204444328943888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,64,0.002445333326856295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,2048,0.015308444698651632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,16384,0.26884977022806805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,128,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,256,0.005036444299750858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,3584,0.007947555846638149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,8192,0.010671999719407825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,128,0.004870222260554631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,128,0.0019386667344305252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,1024,0.012429333395428128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,6144,0.0063528890411059065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,2560,0.004882666799757215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,3072,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,1536,0.015492444237073263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,2560,0.006960888703664144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,16384,0.023200000325838726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,5120,0.016151999433835346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,7168,0.23336709870232475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,6144,0.061624891228146024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,10240,0.027680888772010803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,1536,0.008640888664457533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,3584,0.004915555732117759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,2048,0.03349155518743727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,512,0.006622222148709827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,7168,0.04780800143877665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,6144,0.1306755542755127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,768,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,768,0.016144000821643405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,768,0.02102577851878272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,65536,0.07861155271530151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,16384,0.19068977567884657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,512,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,512,0.07477688789367676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,5120,0.011798222031858234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,5120,0.01000533335738712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,128,0.0021235555824306277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,7168,0.015489778584904142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,7168,0.014779556128713818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,3584,0.00488711107108328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,64,0.006239999913507038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,6144,0.013262222210566202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,32,0.0017608888447284698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,3584,0.020085333122147452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,768,0.004021333323584663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,512,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,512,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,7168,0.010288000106811523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,64,0.002191999927163124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,16384,0.018587556150224473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,4096,0.007916444705592262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,4096,0.007852444218264686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,7168,0.017272889614105225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,3072,0.008000888758235509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,5120,0.008636444807052612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,3584,0.02809244394302368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,7168,0.022720888257026672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,65536,0.05941600269741482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,128,0.005385777602593104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,65536,0.06740088595284356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,16384,0.010064889159467485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,768,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,256,0.004857777721352047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,2048,0.004571555389298333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,512,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,768,0.019732443822754752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,2048,0.02015199926164415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,2048,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,768,0.012776888906955719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,8192,0.014118222726715935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,1536,0.0038311113086011675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,3584,0.011242667006121742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,3584,0.011400889191362592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,6144,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,8192,0.010738666686746808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,3584,0.024715556038750544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,768,0.005881777654091517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,768,0.006317333214812809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,8192,0.013111111190583972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,65536,0.11727200614081489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,1024,0.007303111255168915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,12288,0.025693333811230127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,12288,0.03065599997838338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,10240,0.012103999654452005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,3584,0.009697777529557547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,256,0.00254844439526399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,3072,0.05853866868548923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,16384,0.017620444297790527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,512,0.00254133323000537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,16384,0.05503288904825846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,768,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,2048,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,65536,0.12111200226677789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,2560,0.007296889192528195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,3584,0.032744887802335955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,3072,0.008683555656009251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,768,0.019288000133302476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,32,0.011407111254003314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,8192,0.013066666821638743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,65536,0.07430933581458198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,8192,0.05451288819313049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,8192,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,64,0.01575199928548601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,512,0.0038684445122877755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,8192,0.022506665852334764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,128,0.048545777797698975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,4096,0.007939555578761632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,768,0.025443555580245122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,3072,0.03539289037386576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,65536,0.01721955504682329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,2560,0.006706666615274217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,16384,0.07423022058275011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,5120,0.12211110856797959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,5120,0.008979555633332994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,6144,0.025272000167104933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,2048,0.029502222935358684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,6144,0.016563556260532804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,512,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,16384,0.11089599794811672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,10240,0.012034667034943899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,65536,0.05980000231001112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,3072,0.009680888719028896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,64,0.010686222049925061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,6144,0.029828445778952703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,4096,0.02570755614174737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,65536,0.3342915640936957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,16384,0.019612444771660697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,16384,0.10447200139363606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,1536,0.007294221884674496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,8192,0.00960799968904919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,256,0.0018133332745896445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,32,0.0018515555808941524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,5120,0.017869333426157635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,4096,0.008169777691364288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,32,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,3072,0.07326844665739271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,3584,0.01722400055991279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,1536,0.015831111205948722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,5120,0.014507555299335055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,5120,0.008634666601816813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,4096,0.00443733317984475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,1024,0.014827556080288358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,3072,0.007263110743628607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,3072,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,5120,0.011256888508796692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,3072,0.04638044370545281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,3072,0.011078221930397881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,3584,0.013388444152143268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,12288,0.014055111342006259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,16384,0.04840444525082906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,3072,0.13062132729424372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,1024,0.0035333335399627686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,10240,0.0123848890264829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,16384,0.05163644419776069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,5120,0.009292444421185387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,256,0.007976888782448238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,2048,0.3102702299753825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,2048,0.006588444527652528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,512,0.0052586665583981415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,3072,0.009314667019579146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,256,0.016502221425374348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,3072,0.01658933361371358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,7168,0.07818933327992757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,7168,0.015406222807036506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,2048,0.04058577617009481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,6144,0.12745155228508845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,3584,0.04077688852945963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,768,0.005252444495757421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,512,0.014481777946154276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,10240,0.037442667616738215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,2560,0.00977244476477305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,3584,0.2278266747792562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,16384,0.03348177671432495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,8192,0.015850666496488783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,7168,0.01309866706530253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,7168,0.01756088932355245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,65536,0.029968000120586816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,256,0.005879111174080107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,7168,0.2149813307656182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,3072,0.015533333023389181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,256,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,128,0.010381333529949188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,3584,0.013408000270525614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,3072,0.007231999602582719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,128,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,2560,0.006298666612969504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,256,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,32,0.001820444439848264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,65536,0.1855600012673272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,10240,0.012989333106411828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,3584,0.07694755660163032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,4096,0.013680000272062091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,7168,0.07197066810395983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,12288,0.01824711097611321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,8192,0.019235556324323017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,6144,0.009190222455395592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,3584,0.007607111500369177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,16384,0.015196444259749519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,12288,0.019289778338538278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,5120,0.008464888566070134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,512,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,12288,0.021960000197092693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,7168,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,32,0.0018257778137922287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,16384,0.03523911039034525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,65536,0.41039644347296816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,512,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,1024,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,3584,0.01812266641192966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,3072,0.04850755466355217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,2560,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,64,0.0064399999876817065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,32,0.0021280000607172647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,768,0.01458577811717987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,1536,0.007794666621420119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,16384,0.4979884359571669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,16384,0.22819643550448945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,16384,0.09959910975562201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,128,0.014112888111008538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,8192,0.027104889353116352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,12288,0.028140445550282795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,10240,0.05780444542566935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,1024,0.003671110918124517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,7168,0.14272799756791857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,7168,0.011359111302428775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,6144,0.011905777785513135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,10240,0.016172443826993305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,3072,0.009660444325870937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,32,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,1024,0.005630222045713001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,1024,0.005631999837027655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,768,0.002893333426780171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,128,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,2048,0.07029955254660712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,5120,0.02823288904296027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,1024,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,16384,0.046690666013293795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,2560,0.004550222307443619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,128,0.009655111365848118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,256,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,256,0.014441778262456259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,128,0.004864000197913912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,6144,0.0415013333161672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,6144,0.06225066714816623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,65536,0.08575377861658733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,1024,0.0038515557017591265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,7168,0.00964355551534229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,768,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,128,0.004885333279768626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,768,0.022333333889643352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,6144,0.0839502215385437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,65536,0.05418133404519823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,2560,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,4096,0.022704000274340313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,1536,0.006252444452709622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,65536,1.6750097274780273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,512,0.020968000094095867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,768,0.0063288890653186375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,7168,0.03990311092800564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,512,0.02423733307255639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,3584,0.009002666506502364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,10240,0.028817776176664565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,256,0.011751111182901593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,12288,0.15114043818579778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,256,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,256,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,65536,0.0793164438671536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,12288,0.11414578225877549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,65536,0.019952888290087383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,512,0.0058711109062035876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,256,0.003488000068399641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,8192,0.022984888818528917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,1024,0.06580888562732272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,10240,0.3008106549580892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,4096,0.0079297779334916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,512,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,3072,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,8192,0.01588800052801768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,12288,0.03566044569015503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,16384,0.019664888580640156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,8192,0.06666222545835707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,7168,0.01590755581855774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,768,0.018048000004556444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,65536,0.2531929016113281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,8192,0.015178667174445258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,65536,0.09452711211310492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,6144,0.015524443652894763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,1536,0.06656622224383883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,16384,0.043952001465691455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,1536,0.23513333002726236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,8192,0.033202668031056724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,2048,0.006656888872385025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,128,0.002206222257680363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,128,0.011391999820868174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,6144,0.10407288869222005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,768,0.06956799825032552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,12288,0.021021333005693223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,3584,0.0071937781241205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,16384,0.06916799810197619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,10240,0.01860000027550591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,16384,0.23288710912068686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,256,0.0021048889805873237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,3072,0.013111111190583972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,1536,0.014511111709806653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,128,0.014115555418862237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,1536,0.003563555578390757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,128,0.002529777793420686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,64,0.0068488890926043195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,3584,0.0074471111098925276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,6144,0.02404088940885332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,2560,0.013452444639470843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,10240,0.020868443780475195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,5120,0.012068444656001197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,2560,0.0070168889231152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,32,0.007919111185603673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,12288,0.011722666521867117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,1536,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,768,0.021873777111371357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,4096,0.09536177582210964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,64,0.001973333251145151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,1536,0.004232888834344016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,768,0.002955555501911375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,10240,0.013904000322024027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,2048,0.004600000050332812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,256,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,768,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,5120,0.018910222583346896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,4096,0.009307555854320526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,256,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,4096,0.08121777905358209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,5120,0.040008889304267034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,12288,0.011220444407727985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,3072,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,2560,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,256,0.002465777720014254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,12288,0.12705332703060573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,7168,0.005214222189452913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,1024,0.007669332954618666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,1024,0.0069520001610120135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,128,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,32,0.01610933409796821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,3072,0.07796800136566162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,2048,0.014170666535695394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,2560,0.007228444847795699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,65536,0.056072890758514404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,4096,0.014420444766680399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,1024,0.006229333165619109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,16384,0.039928889936871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,8192,0.009019555317031013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,512,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,4096,0.015486222174432544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,768,0.01293777757220798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,8192,0.0796639985508389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,6144,0.021184888150956895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,2560,0.007143110864692264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,3584,0.008776889079146916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,1024,0.005931555396980709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,64,0.002475555572244856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,6144,0.0273973329199685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,256,0.002512000087234709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,512,0.18226933479309082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,65536,0.6603946685791016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,2048,0.009321778184837764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,512,0.005582222094138463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,4096,0.010648888846238455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,5120,0.012635555532243518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,16384,0.02347822156217363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,5120,0.017610667480362784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,16384,0.0587866637441847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,12288,0.13591289520263672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,3584,0.016151999433835346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,6144,0.004576888763242298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,512,0.004895110925038655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,128,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,2560,0.03211466802491082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,5120,0.016935999194780987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,768,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,4096,0.01018044435315662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,128,0.002190222135848469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,5120,0.018244443668259513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,768,0.017161778277821012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,8192,0.013808888693650564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,2560,0.008451555338170793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,128,0.014422222971916199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,7168,0.028184887435701158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,4096,0.028760890165964764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,128,0.011008000208271874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,3072,0.010666666759385003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,16384,0.08578488561842178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,16384,0.061327113045586475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,8192,0.2683875560760498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,3072,0.008593777815500895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,16384,0.02381422287887997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,16384,0.02568888829814063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,512,0.0032702222880389956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,10240,0.034640000926123724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,512,0.0026906666656335196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,10240,0.04738400048679776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,3584,0.010683555569913654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,7168,0.012474666866991254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,6144,0.013131555583741931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,1536,0.06517244709862603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,256,0.002492444382773505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,7168,0.01777866648303138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,5120,0.05336177680227491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,65536,0.0862942205535041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,1024,0.025250666671329077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,3072,0.005327111317051782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,3584,0.02877955635388692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,1024,0.03661777906947666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,768,0.006566222343179915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,10240,0.009408888717492422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,7168,0.02365866634580824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,10240,0.09356800052854751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,3072,0.008271111382378472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,5120,0.009081777599122789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,3584,0.01701155636045668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,16384,0.02645688917901781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,3584,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,128,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,16384,0.11198488871256511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,128,0.005208000126812193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,3584,0.013903111219406128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,128,0.014313777287801107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,7168,0.06144711044099596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,128,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,3072,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,4096,0.011020444333553314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,1024,0.017303999927308824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,6144,0.004955555415815777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,65536,0.19140622350904676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,1536,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,3584,0.05974311298794217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,1024,0.024028443627887305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,12288,0.019222221440739103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,2048,0.01847555571132236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,4096,0.1559235519833035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,128,0.018901333212852478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,5120,0.019588443968031142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,7168,0.03871644536654154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,5120,0.025225778420766194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,10240,0.03829511006673177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,10240,0.029134223858515423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,10240,0.02289688918325636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,3072,0.027729777826203242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,6144,0.02079288827048408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,128,0.004600000050332812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,32,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,768,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,8192,0.014119111829333834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,2560,0.007159999675220913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,2560,0.0229759994480345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,1536,0.006630222416586346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,1536,0.01790488925245073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,7168,0.009943111075295342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,2560,0.01108444482088089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,512,0.01777066621515486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,512,0.01311555587583118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,3072,0.01518133282661438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,768,0.003555555724435382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,256,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,2560,0.012632889052232107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,1536,0.046682665745417275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,768,0.02032711108525594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,4096,0.3389466603597005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,2560,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,8192,0.014581332604090372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,768,0.003526222374704149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,1536,0.006585777633719974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,3072,0.01479200025399526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,128,0.004867555366622077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,7168,0.14032177130381265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,65536,0.06461777951982287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,2048,0.014535999960369535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,128,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,2560,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,1536,0.005738666488064661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,6144,0.009679111341635386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,128,0.011760888828171624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,256,0.002193777718477779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,2048,0.006605333338181178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,768,0.008619555168681675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,128,0.006600889066855113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,4096,0.08887377712461685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,1536,0.01756355497572157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,512,0.014151111245155334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,1024,0.025029333101378545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,768,0.0141004439857271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,768,0.0028515555378463534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,1536,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,6144,0.009352888498041365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,768,0.006069333189063602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,6144,0.025757332642873127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,3072,0.1812080012427436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,3072,0.013746666411558786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,65536,0.16375288698408338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,12288,0.02236977716286977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,12288,0.02788089050187005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,4096,0.01000355515215132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,256,0.00250133333934678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,256,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,6144,0.027837332752015855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,3584,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,1536,0.01347555551264021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,3072,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,1024,0.02072533302836948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,768,0.033265779415766396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,8192,0.010046222143703038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,16384,0.01737244427204132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,6144,0.06079733371734619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,16384,0.02456444501876831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,512,0.022052443689770166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,5120,0.39330045382181805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,5120,0.008612444831265343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,6144,0.04660977919896444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,512,0.0029155556112527847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,7168,0.012262221839692859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,6144,0.08603644371032715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,768,0.006048888795905643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,65536,0.08759466807047527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,2048,0.006299555715587404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,2560,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,1024,0.002830222249031067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,32,0.0029075555503368378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,5120,0.0069413334131240845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,256,0.005331555588377847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,6144,0.01385244478782018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,2048,0.006245333287451003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,10240,0.030064000023735896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,1024,0.003922666526503033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,32,0.004285333471165763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,7168,0.09004444546169704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,10240,0.017289777596791584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,5120,0.026791999737421673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,8192,0.024654222859276667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,5120,0.01865244408448537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,65536,0.029303110308117334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,10240,0.04284088810284933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,2048,0.006281777802440856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,3584,0.008074666890833113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,1536,0.018538667096032035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,65536,0.08370400137371486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,8192,0.020816889074113634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,3072,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,12288,0.05257155497868856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,2560,0.026033777329656813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,256,0.014485332700941296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,10240,0.022277333670192297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,1024,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,5120,0.086681776576572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,7168,1.3994604746500652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,512,0.005988444305128521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,128,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,3584,0.009780444204807281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,1024,0.02872177627351549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,768,0.014152000347773233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,256,0.004933333231343163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,10240,0.3998844358656142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,4096,0.008700444466537898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,3072,0.04722222354676989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,12288,0.07254488600624932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,6144,0.01555199921131134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,3072,0.007278222176763747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,768,0.005274666680230035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,6144,1.9392221238878038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,3584,0.026499556170569524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,2048,0.012713777522246042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,768,0.0032195556494924757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,12288,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,768,0.005583999885453119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,768,0.005513777749405966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,10240,0.024009777439965144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,128,0.010048888623714447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,12288,0.07849511173036364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,3584,0.008778666456540426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,2048,0.0277511113219791
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,8192,0.011036444041464063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,3584,0.03896888759401109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,2560,0.008056888977686564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,6144,0.022436444958051045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,3072,0.007341333561473423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,4096,0.018414222531848483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,64,0.010355555348926121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,6144,0.03166044420666165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,1536,0.011343999869293638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,2560,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,3584,0.007914666500356462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,12288,0.006996444529957241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,7168,0.03640888796912299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,768,0.10848533444934422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,32,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,256,0.009014222357008193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,6144,0.01754044493039449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,64,0.005621333503060871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,5120,0.028103109863069322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,5120,0.008832889298597971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,256,0.0021884443445338146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,16384,0.023027555810080633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,128,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,16384,0.0630542238553365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,65536,0.1748382250467936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,3072,0.04164800047874451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,16384,0.032073779238594904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,7168,0.036610666248533465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,3584,0.005583999885453119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,7168,0.015107555521859063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,512,0.005290666802061929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,12288,0.029861334297392104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,6144,0.03416444526778327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,6144,0.04618933465745714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,128,0.009500444763236577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,4096,0.01110755569405026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,12288,0.012786666552225748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,256,0.011758222348160215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,3072,0.01313244468635983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,1024,0.005923555543025334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,3584,0.004899555610285865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,16384,0.11932178338368733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,512,0.003371555565132035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,6144,0.012446222205956778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,128,0.004592000196377437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,2560,0.015623999966515435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,2048,0.00637511122557852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,128,0.024007999234729346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,128,0.00488711107108328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,10240,0.03440977798567878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,3072,0.012325333224402534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,512,0.00564177789621883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,4096,0.02053155501683553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,4096,0.01554666625128852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,3072,0.00833155545923445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,65536,0.123233781920539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,4096,0.010430222584141625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,2560,0.06710666418075562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,256,0.005626666463083691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,16384,0.05017422305213081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,3072,0.007273777491516537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,16384,0.06876533561282687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,1536,0.0366559988922543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,4096,0.09064000182681614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,1024,0.0038853333228164246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,1536,0.012395555774370829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,1024,0.005962666538026597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,1536,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,512,0.01479555500878228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,65536,0.03638933433426751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,2048,0.020991111795107525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,2560,0.009678222239017487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,768,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,3072,0.03728177812364366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,4096,0.007908444437715743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,6144,0.008996444443861643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,768,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,10240,0.04605244596799215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,16384,0.1678008900748359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,32,0.004915555732117759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,65536,0.05727377865049574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,256,0.004979555391603046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,10240,0.02502311103873783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,6144,0.020597333709398907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,1536,0.019897777173254225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,768,0.0028151110228565005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,3584,0.01644622286160787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,512,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,128,0.014463999205165438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,128,0.00591111100382275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,2560,0.006950221955776215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,1536,0.10444355673260158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,10240,0.026439110438028973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,256,0.004958222309748332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,128,0.013414222333166333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,12288,0.013818666338920593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,3072,0.023080888721677992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,5120,0.036278221342298723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,5120,0.010423111418883005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,3584,0.014141332772043018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,1024,0.03565244542227851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,3072,0.013451555536852943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,32,0.002492444382773505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,7168,0.009296000003814697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,32,0.005664000080691443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,12288,0.020928000410397846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,16384,0.07578933238983154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,16384,0.007644444704055786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,65536,0.08246488703621759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,6144,0.018252443936136033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,65536,0.036166220903396606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,12288,0.08859555588828193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,16384,0.15617067284054226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,3072,0.014689778288205465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,2560,0.08296266529295179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,2048,0.007654222349325816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,6144,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,12288,0.019919999771647982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,5120,0.012294222083356647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,256,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,6144,0.03558489018016391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,5120,0.02257866660753886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,768,0.014459555347760519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,10240,0.10111466381284927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,12288,0.22486755583021376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,8192,0.011144888897736868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,128,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,128,0.004960888789759742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,65536,0.0680604444609748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,7168,0.009917333722114563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,768,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,128,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,6144,0.009314667019579146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,7168,0.021433777279324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,4096,0.03086400032043457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,4096,0.017667555146747164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,1024,0.009732444253232744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,65536,0.1063422229554918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,12288,0.017197334104114108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,128,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,64,0.002083555484811465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,512,0.00259911103381051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,16384,0.016855110724767048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,768,0.010458666417333813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,5120,0.04020266731580099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,5120,0.07924444145626493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,5120,0.008354666332403818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,3072,0.006649777707126405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,256,0.014827556080288358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,256,0.002899555489420891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,5120,0.018782221608691745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,768,0.0063742221229606206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,1536,0.049154665735032826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,10240,0.023748444186316595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,3584,0.03231733375125461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,6144,0.10021333562003241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,1536,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,65536,0.04741955465740628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,128,0.002642666714058982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,3584,0.23092712296379936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,2048,0.004253333227501975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,12288,0.014013333453072442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,768,0.007889777421951294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,768,0.0380311111609141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,16384,0.03502755694919162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,65536,0.18866399923960367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,256,0.004879999905824661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,768,0.01647733317481147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,10240,0.13080178366767034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,1536,0.006670222100284364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,1536,0.0062977779242727495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,3584,0.04755022128423055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,3072,0.4817493226793077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,256,0.014870221416155497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,6144,0.01792622274822659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,512,0.005912000106440649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,10240,0.10990489191479152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,7168,0.011117333339320289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,12288,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,512,0.0028533333291610083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,2048,0.010117332968446944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,12288,0.015786666009161208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,32,0.01717333330048455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,3584,0.02260444396071964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,65536,0.282367123497857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,3584,0.0069875551594628235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,2048,0.02032533288002014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,5120,0.00904977818330129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,16384,0.028751999139785767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,128,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,768,0.0058746664888328975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,6144,0.007984889050324758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,64,0.002171555534005165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,3584,0.007600888609886169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,8192,0.018538667096032035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,1536,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,10240,0.01699022286468082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,64,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,128,0.002132444539003902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,16384,0.015150222513410779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,16384,0.13989422056410047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,128,0.007285333342022366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,64,0.00554577757914861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,128,0.012443555725945367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,2048,0.02010222276051839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,1536,0.006262222097979651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,512,0.01685333251953125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,3584,0.018927999668651156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,128,0.014163555370436775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,6144,0.054119110107421875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,7168,0.012069332930776807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,512,0.011727111207114326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,2560,0.004544888933499654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,6144,0.03395466672049628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,65536,0.05752888653013441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,64,0.027059555053710938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,128,0.002492444382773505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,2048,0.011431111229790581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,6144,0.02314222190115187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,16384,0.024153777294688757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,1024,0.0031582222630580268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,12288,0.04180533356136746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,2560,0.00943022221326828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,256,0.01275111072593265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,256,0.005929777605666055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,12288,0.029772443903817072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,7168,0.014153778553009033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,65536,0.15020178423987493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,1536,0.06807200113932292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,3584,0.02537333303027683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,128,0.004965333475006951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,65536,0.05669777923160129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,8192,0.013184888495339287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,1024,0.013748444616794586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,4096,0.01792888840039571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,2560,0.008343111309740279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,7168,0.022856000396940444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,65536,3.4158852895100913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,10240,0.040906667709350586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,64,0.002143111079931259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,5120,0.09446489148669773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,512,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,8192,0.028451555305057104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,16384,0.026174222429593403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,7168,0.025264889001846313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,65536,0.05338311195373535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,6144,0.18558932675255668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,1024,0.004231111043029361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,12288,0.03943466809060838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,4096,0.019129777948061626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,7168,0.02304977840847439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,2048,0.004553777890072929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,65536,4.666650560167101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,6144,0.009698666632175446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,1536,0.013798221945762634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,7168,0.05948266718122694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,12288,0.2988319926791721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,512,0.002793777734041214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,65536,0.021319111188252766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,6144,0.023006222314304773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,256,0.014752000570297241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,1024,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,1024,0.012536889149083031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,8192,0.07517421907848783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,12288,0.029401779174804688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,768,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,128,0.013748444616794586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,3072,0.013083555632167392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,3072,0.02401599950260586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,64,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,10240,0.025275554921891954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,4096,0.008958222137557136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,16384,0.03417155477735732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,7168,0.009312888814343346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,2048,0.012543111211723752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,5120,0.012424888710180918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,5120,0.012404444317022959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,128,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,12288,0.11190577348073323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,4096,0.07071733474731445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,6144,0.0073902226156658595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,10240,0.022724444667498272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,16384,0.03507466779814826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,16384,0.20528799957699248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,5120,0.008995555341243744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,6144,0.009202666580677032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,2560,0.012410666379663678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,3584,0.021310221817758348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,2560,0.014098667436175875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,768,0.016547555724779766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,2048,0.0041893331540955436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,7168,0.011056889262464313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,128,0.01035911093155543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,2560,0.007910221815109253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,1024,0.027355555031034682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,5120,0.012412444584899478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,16384,0.005589333259397083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,4096,0.008018666671382057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,6144,0.00924622184700436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,128,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,64,0.0018097776919603348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,6144,0.009384888741705153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,3584,0.007839110990365347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,1024,0.02836444311671787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,512,0.0034977777136696708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,5120,0.017785777648289997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,1024,0.015997333659066092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,6144,0.29737067222595215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,3584,0.009392889009581672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,2048,0.006597333484225803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,3584,0.09275022480222915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,256,0.00498933345079422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,10240,0.023445333043734234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,65536,0.1182817750506931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,1536,0.015110222829712762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,10240,0.02404444416364034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,7168,0.00997155573632982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,3072,0.02095288866096073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,2048,0.013551111022631327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,768,0.005537777725193236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,5120,0.02243288854757945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,768,0.006958222223652734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,7168,0.007306666837798224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,2048,0.00699733363257514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,128,0.007283555964628856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,7168,0.009673777553770278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,3072,0.02539199921819899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,512,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,5120,0.014124444789356656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,3072,0.009336888790130615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,64,0.0021333334346612296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,65536,0.11735555860731338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,8192,0.031662222411897444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,512,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,8192,0.012521777715947894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,3584,0.026100445124838088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,512,0.0028968888024489083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,6144,0.009094222552246517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,5120,0.013368888861603208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,512,0.019368888603316415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,1024,0.006596444381607904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,7168,0.014827556080288358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,12288,0.03453333179155985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,64,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,3072,0.006948444578382704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,64,0.00999466660949919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,10240,0.014875556031862894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,2048,0.006687110910813014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,768,0.0028311111446883944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,65536,0.08911466598510742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,65536,0.07831199963887532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,2560,0.01611911091539595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,12288,0.05801244576772054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,3072,0.007040888898902469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,128,0.0063271112740039825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,7168,0.009525333013799455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,2560,0.008180444439252218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,6144,0.02873777680926853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,3072,0.007276444799370236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,1536,0.007558222446176741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,512,0.022709333234363135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,5120,0.015829333000712924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,2048,0.007897777689827813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,8192,0.015452444553375244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,12288,0.0207013338804245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,512,0.015847999188635085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,6144,0.01182666669289271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,10240,0.015811555915408664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,3584,0.015472888946533203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,8192,0.04522222280502319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,3584,0.009347555538018545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,7168,0.03679822219742669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,256,0.016750221451123554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,32,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,64,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,3072,0.01516177753607432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,16384,0.04085777865515815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,256,0.002201777779393726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,4096,0.09563733471764459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,768,0.006254222244024277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,7168,0.018988443745507132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,256,0.0039022221333450745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,8192,0.02373955481582218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,2560,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,7168,0.02141511109140184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,2560,0.014122666584120857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,16384,1.699228392706977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,65536,0.11418399545881484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,512,0.0032088889016045462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,12288,0.040080888403786555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,4096,0.04589688777923584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,16384,0.07719644572999743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,4096,0.024254222710927326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,768,0.002834666727317704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,3584,0.02399377855989668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,6144,0.045591112640168935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,4096,0.015469332536061605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,512,0.015480000111791821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,65536,0.4021351072523329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,16384,0.04743999905056424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,8192,0.01722400055991279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,256,0.005295111073387994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,256,0.00489688871635331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,65536,1.036985821194119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,8192,0.024495111571417913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,2560,0.01440355512830946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,256,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,512,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,6144,0.011332444846630096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,1536,0.007650666766696506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,8192,0.015276443627145557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,128,0.00598933340774642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,3072,0.0047031111187405055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,128,0.0021991110924217436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,5120,0.008634666601816813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,256,0.006225777582989798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,8192,0.014484445254007975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,8192,0.04721600148412916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,16384,0.08599199851353963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,16384,0.03138577938079834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,1024,0.014473777678277759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,64,0.005877333382765452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,32,0.001840888833006223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,3072,0.007198221981525421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,2048,0.11043110820982192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,65536,0.00869599978129069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,128,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,16384,0.023253333237436082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,65536,0.09459822045432197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,4096,0.01401600076092614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,2560,0.009783999787436591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,64,0.005615111026499007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,64,0.003979555434650845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,512,0.0029173334025674393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,12288,0.05935555696487427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,6144,0.02730400032467312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,5120,0.04333688815434774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,3584,0.014883556299739413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,768,0.00562755556570159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,10240,0.02031822171476152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,7168,0.14963377846611872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,12288,0.03899288839764065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,5120,0.024254222710927326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,2560,0.00684533350997501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,1536,0.003938666648334927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,32,0.0028906667398081887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,2048,0.018623110320832994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,12288,0.013429332938459186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,6144,0.012349333200189801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,6144,0.012124444047609964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,8192,0.13934755325317383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,16384,0.022072000635994807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,32,0.006240889016124938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,2560,0.02260977857642704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,256,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,10240,0.01071200023094813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,6144,0.00527022240890397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,64,0.005938666562239329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,128,0.0021760000122918025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,10240,0.37688178486294216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,512,0.003198222153716617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,2560,0.012776888906955719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,768,0.009313777916961247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,2560,0.011702222128709158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,3584,0.020000000794728596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,3584,0.011376000112957425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,768,0.005538666827811136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,5120,0.00866844422287411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,10240,0.06852710909313627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,2560,0.006631111105283101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,8192,0.45563909742567277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,1536,0.00600000015563435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,32,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,32,0.0029102222373088202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,3584,0.007992000215583378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,6144,0.01074577785200543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,3072,0.008985777695973715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,12288,0.02236355510022905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,3584,0.010196444061067369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,4096,0.016591111818949383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,1024,0.0032142222755485107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,256,0.0034799998005231223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,128,0.006586666736337874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,32,0.002785777880085839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,10240,0.02554133368863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,512,0.051925334665510386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,3584,0.08899822500016953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,12288,0.02735911144150628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,6144,0.03277511066860623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,768,0.005592000153329637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,128,0.006296000132958095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,2560,0.011807999677128263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,32,0.0021217777911159727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,3584,0.026030222574869793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,7168,0.009355555805895064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,7168,0.014797333214018079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,2048,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,16384,0.026408889227443274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,3072,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,6144,0.015997333659066092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,10240,0.005331555588377847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,3584,0.008288000192907121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,4096,0.007617777420414819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,65536,0.122524446911282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,6144,0.10186577505535549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,1024,0.0038408889538711975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,512,0.00490133340160052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,6144,0.011718221836619906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,32,0.01238933371173011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,10240,0.02380977736579047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,8192,0.4597439765930176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,16384,0.034352888663609825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,2048,0.012727110750145383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,1536,0.02128711177243127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,65536,0.2765013376871745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,1024,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,64,0.006380444599522485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,2560,0.006908444480763541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,128,0.013802666631009845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,6144,0.014488889111412896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,1536,0.01348000019788742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,3072,0.024680889315075342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,512,0.002551111082235972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,6144,0.01647111111217075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,6144,0.2833342287275526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,10240,0.014789332946141561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,8192,0.017521777086787753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,7168,0.00961599995692571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,3072,0.026114667455355327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,256,0.026727111803160772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,256,0.017622222503026325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,1024,0.021684444612926904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,12288,0.8156835767957898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,2560,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,12288,0.03401688734690348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,2048,0.018970666660202872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,1024,0.017613333132531907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,32,0.0029208889851967492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,3584,0.007506666911972894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,2560,0.009443555441167619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,512,0.006020444548792309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,16384,0.02529688841766781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,1536,0.01184266640080346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,3584,0.004296888907750447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,16384,0.023401778605249193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,65536,0.05189866820971171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,3072,0.11038133170869614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,16384,0.02017866737312741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,10240,0.08215288983450995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,5120,3.2836551666259766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,3584,0.014335999886194864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,3584,0.01612622208065457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,5120,0.00902844468752543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,8192,0.01035733355416192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,1024,0.009013333254390294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,16384,0.042952001094818115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,10240,0.036666668123669095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,10240,0.012470222181744046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,256,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,3072,0.017893332574102614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,2048,0.24604977501763237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,3072,0.02201333310869005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,4096,0.029494222667482164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,512,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,768,0.49457597732543945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,7168,0.009018667042255402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,2560,0.015602666470739575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,5120,0.05044088761011759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,4096,0.005895111295912001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,512,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,1536,0.01925066610177358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,10240,0.18534489472707114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,3584,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,3072,0.008551110823949179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,1024,0.01201244443655014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,768,0.2530328962537977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,1024,0.013430222041077085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,3584,0.01610488858487871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,64,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,512,0.0025315555847353404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,10240,0.008072888685597314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,3584,0.008047111332416534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,7168,0.019389332996474374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,1024,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,5120,0.008355555435021719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,768,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,2048,0.006263111200597551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,4096,0.02235288918018341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,3584,0.007285333342022366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,2048,0.006591111007663939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,10240,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,65536,0.06485155555937025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,4096,0.011761777930789523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,65536,0.11233688725365533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,128,0.004952888935804367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,7168,0.013452444639470843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,128,0.002181333386235767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,1536,0.016162667009565566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,1024,0.003157333367400699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,2048,0.006440889090299606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,5120,0.01316088851955202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,65536,0.022794667217466567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,5120,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,3072,0.01036800030204985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,128,0.0021066665649414062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,16384,0.08703911304473877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,512,0.012413333687517377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,6144,0.15146133634779188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,1024,0.01555288831392924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,768,0.006272000157170826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,2048,0.006640888750553131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,3072,0.01072799993885888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,7168,0.012118221984969245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,2048,0.014771555860837301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,5120,0.03364355696572198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,512,0.00589155571328269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,65536,0.08349866337246364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,2048,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,256,0.005024888863166173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,8192,0.021141333712471858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,7168,0.027444443768925134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,1024,0.013448889056841532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,8192,0.005610666755172942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,16384,0.04085155659251743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,1536,0.009455111291673448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,8192,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,3584,0.05837777588102552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,512,0.012503110700183444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,4096,0.01018133345577452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,4096,0.05072355601522657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,64,0.0018435555199782054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,256,0.014675555957688225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,1536,0.003855111284388436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,4096,0.0041999999019834726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,2048,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,8192,0.07095022333992852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,6144,0.2193982203801473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,768,0.016843555702103507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,512,0.012221333053376941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,3584,0.012272000312805176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,6144,0.014291556345091926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,64,0.012079111403889127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,2048,0.004351110921965705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,16384,0.06731910838021173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,65536,0.07570133606592815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,128,0.006248000181383557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,65536,0.04656622144911024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,10240,0.26834310425652397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,12288,0.029267556137508813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,128,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,16384,0.023641778363121882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,10240,0.020283554991086323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,768,0.03377422359254625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,1536,0.006237333433495627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,4096,0.021741333934995864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,10240,0.022351110974947613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,12288,0.2860026624467638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,7168,0.009687110781669617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,1024,0.007287999822033777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,768,0.0031911111954185697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,5120,0.035904887649748064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,256,0.02600266701645321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,3584,0.005504888792832692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,64,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,6144,0.010778666370444827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,5120,0.021576000584496394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,3584,0.1671004427803887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,512,0.02123822271823883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,16384,0.027705777022573683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,2560,0.008984888593355814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,7168,0.1341315507888794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,6144,0.015154666370815702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,128,0.004906666775544484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,3072,0.05524888965818617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,5120,0.01869777838389079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,4096,0.021168889270888433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,4096,0.030784888399971858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,1024,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,3072,0.007616889145639207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,3584,0.011100444528791638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,8192,0.00794400026400884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,5120,0.008626666333940294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,5120,0.028016000986099243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,6144,0.017288888494173687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,128,0.008681777450773451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,3072,0.009672889278994666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,2560,0.01661066710948944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,8192,0.314862224790785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,6144,0.019965334071053397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,1024,0.0038897775941424896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,768,0.021432888176706102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,256,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,768,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,65536,0.32745154698689777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,1536,0.020908445119857788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,7168,0.019351111518012155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,5120,0.02661955522166358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,256,0.004939555707905027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,2560,0.007236444287829929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,16384,0.02328355610370636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,256,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,7168,0.010656000011497073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,2048,0.011750222080283694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,7168,0.009988444546858469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,1536,0.013412444127930535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,768,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,65536,0.057293335596720375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,3072,0.007248000138335758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,768,0.007732444339328342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,8192,0.01036533299419615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,6144,0.03795199924045139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,512,0.0063866666621632045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,1536,0.0042453333735466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,64,0.007292444507280986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,10240,0.015806222955385845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,128,0.004922666483455234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,64,0.010384889112578498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,8192,0.02986488739649455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,65536,0.01720266706413693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,10240,0.02596355477968852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,10240,0.012481778032249875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,2560,0.02016444504261017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,256,0.012134222520722283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,5120,0.237119992574056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,12288,0.013287111288971372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,128,0.014428445034556918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,2560,0.013079110946920184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,512,0.01237777786122428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,512,0.013481777575280933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,5120,0.005990222096443176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,5120,0.019574221637513902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,7168,0.009658666948477427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,6144,0.08775111039479573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,768,0.017886221408843994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,64,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,3584,0.005884444548024072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,32,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,6144,0.007978666987684038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,12288,0.06740622387992011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,256,0.01615111033121745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,4096,0.08506666951709324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,4096,0.014118222726715935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,2048,0.008997333546479544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,2048,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,3584,0.041165331999460854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,8192,0.01992177797688378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,512,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,3584,0.011705777711338468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,256,0.0147724449634552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,3584,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,2560,0.06152088774575127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,768,0.0028764444092909494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,32,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,1024,0.007241778075695038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,64,0.0017831110292010838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,128,0.013779555757840475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,2560,0.006656000183688269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,1024,0.03180088930659824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,32,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,10240,0.012060444388124677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,4096,0.056999113824632436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,256,0.0022320000247822865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,32,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,2048,0.006582222051090664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,768,0.002971555623743269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,4096,0.09671555625067817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,16384,0.03232889042960273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,3072,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,768,0.00554044461912579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,65536,0.08315200275844999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,6144,0.02955289019478692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,65536,0.16625244087643093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,4096,0.008048888709810045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,6144,0.013883555928866068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,256,0.014519999424616495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,7168,0.07845866680145264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,256,0.004619555754794015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,6144,0.009385777844323052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,512,0.013807111316257052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,8192,0.04267911116282145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,5120,0.0076044441925154785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,64,0.0038906666967603895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,16384,0.010058666268984476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,2048,0.015806222955385845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,10240,0.00692622239391009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,10240,0.016528000434239704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,3072,0.007599111232492659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,3584,0.02860622273551093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,4096,0.007711110843552484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,1536,0.021781333618693884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,10240,0.023494222097926672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,6144,0.19985332753923205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,4096,0.014165333575672574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,16384,0.015138667490747241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,256,0.01683555543422699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,6144,0.02541422181659275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,512,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,3584,0.007995555798212687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,10240,0.02611200014750163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,65536,0.121979554494222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,3072,0.010044444766309526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,7168,0.18775555822584364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,768,0.014837332897716098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,2560,0.015170666906568738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,128,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,2048,0.0689200030432807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,12288,0.029853334029515583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,10240,0.013529777526855469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,10240,0.012736000120639801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,1024,0.01276266657643848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,1024,0.0465280016263326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,512,0.0032222221295038858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,7168,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,768,0.005582222094138463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,1024,0.005639111002286275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,2560,0.007311999797821045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,768,0.049548443820741445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,32,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,2048,0.02159022291501363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,512,0.015154666370815702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,7168,0.009402666654851701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,32,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,3072,0.01889333294497596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,65536,0.075655996799469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,16384,0.04425155454211765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,2048,0.006648889018429651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,1536,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,10240,0.016493333710564505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,6144,0.03863288958867391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,16384,0.016846223009957206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,5120,0.010669333239396414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,512,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,3072,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,3072,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,512,0.003165333428316646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,3584,0.009280888570679558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,10240,0.020612445142534044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,3584,0.053787555959489614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,10240,0.16673244370354545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,1536,0.011399110986126794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,8192,0.015801777442296345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,10240,0.012689777546458773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,12288,1.1058657964070637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,3072,0.006285333385070165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,3072,0.020664888951513503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,65536,0.04728800058364868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,16384,0.04349155558480156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,512,0.007287110719415877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,2560,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,10240,0.024655110306209985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,5120,0.01930488811598884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,64,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,128,0.006587555425034628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,16384,0.02302933401531643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,2048,0.006613333192136552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,2048,0.006308444258239534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,7168,0.1376604504055447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,7168,0.30274399121602374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,2048,0.020959999826219346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,5120,0.006288888967699475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,1024,0.006215999937719769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,768,0.010445333189434476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,256,0.002459555657373534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,8192,0.0955502192179362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,5120,0.008596444295512306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,512,0.005533333453867171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,512,0.004552888787455029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,10240,0.019527999891175162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,1536,0.0034728890491856467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,3072,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,7168,0.30236710442437065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,128,0.020628444022602506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,256,0.0038720000949170855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,3072,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,10240,0.019660444723235238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,256,0.006650666809744305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,768,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,128,0.08030666907628377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,4096,0.008277333445019191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,16384,0.033554666572146945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,3584,0.004230222354332606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,6144,0.013265777793195514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,128,0.00518222235971027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,16384,0.013707555830478668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,12288,0.13828799459669325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,512,0.0024702221983008915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,64,0.007652444144090016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,64,0.002453333387772242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,64,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,8192,0.05537422166930305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,2560,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,128,0.004912000149488449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,12288,0.025723555021815833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,6144,0.013239111337396832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,768,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,512,0.01649688846535153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,10240,0.036631110641691424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,64,0.03621777892112732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,16384,0.029493331909179688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,256,0.002485333424475458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,32,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,5120,0.050606220960617065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,2048,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,6144,0.018082666728231642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,4096,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,16384,0.04913333389494154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,16384,0.022833777798546687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,5120,0.013088888592190213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,5120,0.029704888661702473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,512,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,1024,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,64,0.004875555634498596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,3072,0.004553777890072929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,65536,2.9278498755560984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,2560,0.006606222026877933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,8192,0.10747022098965114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,512,0.006665777828958299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,768,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,8192,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,6144,0.009823111196359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,768,0.018202667435010273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,3584,0.012154666913880242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,64,0.0018719999740521114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,4096,0.007666666474607255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,2048,0.007810666329330868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,3072,0.007914666500356462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,64,0.0025395556456512874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,65536,0.23884712325202095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,6144,0.03836977812978957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,12288,0.9596000247531467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,16384,0.09700800312889947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,1024,0.007334222396214803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,512,0.01650399963061015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,1536,0.005517333332035277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,8192,0.10039022233751084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,3072,0.011032889286677042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,16384,0.029558221499125164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,1536,0.015166223049163818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,4096,0.01076533314254549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,65536,0.07335555553436279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,7168,0.0266942216290368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,1536,0.005950222412745158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,3072,0.008061333662933773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,1536,0.003875555677546395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,3584,0.07031110922495525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,2560,0.0069955554273393415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,7168,0.009992000129487779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,1536,0.006252444452709622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,7168,0.03403910994529724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,2560,0.015785778562227886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,5120,0.031062222189373438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,3072,0.545538690355089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,512,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,3072,0.017867555220921833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,64,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,4096,0.032418668270111084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,128,0.014429334137174817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,16384,0.025472889343897503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,128,0.020615110794703167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,5120,0.012207110722859701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,65536,0.04456888967090183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,16384,0.03219288918707106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,16384,0.024671999944580927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,1024,0.023207111491097346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,128,0.012063999970753988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,65536,0.07685066594017877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,768,0.02198311189810435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,8192,0.008720888859695857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,2560,0.018719111879666645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,128,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,1536,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,1536,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,3072,0.007146666447321574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,10240,0.023881778120994568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,1536,0.03700088792377048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,12288,0.00527555536892679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,10240,0.01384888920519087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,12288,0.004880889008442561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,12288,0.02538577715555827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,4096,0.01495822270711263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,1536,0.043613334496816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,4096,0.020446222689416673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,10240,0.10861244466569688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,32,0.0018373332503769135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,10240,0.00795022232664956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,65536,0.07451733615663317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,16384,0.0761573314666748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,3072,0.025655110677083332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,768,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,256,0.006615110983451207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,1024,0.007296889192528195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,2560,0.005583999885453119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,2560,0.00665866666369968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,3072,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,2048,0.00793866647614373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,65536,0.6084853278266059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,12288,0.18815999560885957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,512,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,1024,0.006456888798210356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,3072,0.021543110410372417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,256,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,12288,0.012741333080662621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,128,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,7168,0.011733333269755045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,2048,0.020263999700546265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,4096,0.0064693333374129395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,3072,0.010464888479974536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,256,0.006623999940024481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,32,0.005609777652555042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,3072,0.010055999788973067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,3584,0.020829333199395072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,768,0.006761777732107375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,3072,0.007271111011505127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,1024,0.013942221800486246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,512,0.012286221815480126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,128,0.006320000108745363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,256,0.002502222235004107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,3072,0.007282666862010956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,10240,0.012069332930776807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,1536,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,768,0.005620444400442972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,256,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,10240,0.033033778270085655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,16384,0.027140445179409448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,8192,0.05554133322503832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,7168,0.01828266680240631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,512,0.008257778154479133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,65536,0.2590586609310574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,2560,0.025969778498013813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,3072,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,1024,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,16384,0.022991110881169636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,6144,0.00981422182586458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,4096,0.007988444632954067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,6144,0.03365155392222934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,32,0.0021271111650599376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,3072,0.019349333312776353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,16384,0.04444977641105652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,3072,0.0148062225845125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,8192,0.014280888769361707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,7168,0.020522667302025687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,5120,0.02999733222855462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,256,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,1024,0.00703288863102595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,65536,0.06135111384921604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,1024,0.008473777936564552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,1536,0.02646222213904063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,6144,0.013773333695199756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,6144,0.012814222110642327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,16384,0.01978755493958791
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,4096,0.05946399768193563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,64,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,128,0.013646221823162503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,12288,0.044679999351501465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,8192,0.013417777915795645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,768,0.004022222426202563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,128,0.01795111099878947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,8192,0.03085244364208645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,64,0.005869333528810077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,10240,0.05284088850021362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,128,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,65536,0.06831022103627522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,256,0.0058062221441004015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,65536,0.21131822797987196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,7168,0.020470221837361652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,6144,0.011792889071835412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,1024,0.005439110928111606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,3072,0.007323555648326874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,6144,0.020922667450375024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,65536,2.3121280670166016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,7168,0.02369955513212416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,3072,0.009661333428488838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,2560,0.007260444263617198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,4096,0.01681422193845113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,7168,0.010040000081062317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,2048,0.014467555615637036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,10240,0.32506577173868817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,10240,0.2977093325720893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,1024,0.005924444645643234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,10240,0.11190756162007649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,768,0.003929777691761653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,32,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,768,0.0029111111329661477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,7168,0.009323555562231276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,2048,0.0069466663731469055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,768,0.030493332280053034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,64,0.0063662222690052455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,8192,0.02717155549261305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,8192,0.02302933401531643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,65536,0.1839368873172336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,65536,0.40717776616414386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,32,0.0018800000349680583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,32,0.018223111828168232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,32,0.009329777624871995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,3584,0.014779556128713818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,128,0.006937777830494775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,3584,0.1330746677186754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,1536,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,2048,0.011713777979214987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,128,0.005385777602593104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,7168,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,32,0.012131555212868584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,65536,0.009128889275921715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,3072,0.050753778881496854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,8192,0.02868888775507609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,16384,0.019463999403847587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,5120,0.011342222491900126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,1536,0.013754666679435305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,6144,0.015992000699043274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,128,0.018548443913459778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,2560,0.005277333160241445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,768,0.006960888703664144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,3584,0.00794222205877304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,512,0.0064426664676931165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,512,0.0090844440791342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,256,0.0032115555885765287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,10240,0.03195822238922119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,2048,0.006801777829726537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,1536,0.006228444476922353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,7168,0.029318223396937054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,10240,0.3315315511491564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,16384,0.16908800601959229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,2560,0.006903999795516332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,64,0.00656444455186526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,768,0.0025902222841978073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,8192,0.1333964400821262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,10240,0.016929777132140267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,2560,0.004640000147951974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,1024,0.006326222171386083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,65536,0.41032889154222274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,10240,0.016728000508414376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,64,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,7168,0.025037333369255066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,12288,0.01549600064754486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,65536,0.10404711299472386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,4096,0.00795199970404307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,1536,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,256,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,12288,0.024433778391944036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,7168,0.17735732926262748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,5120,0.012449777788586087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,10240,0.017795556121402316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,2048,0.02772444486618042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,8192,0.12443555725945367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,32,0.005541333307822545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,128,0.0021102221475707158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,16384,0.030248890320460003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,65536,0.006295111030340195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,1024,0.0035004446076022256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,32,0.005867555737495422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,3584,0.009723555710580613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,64,0.006302222195598815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,8192,0.01070666644308302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,6144,0.013321778012646569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,8192,0.026404443714353774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,65536,0.08559022347132365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,12288,0.020642666353119742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,64,0.005522666705979242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,768,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,2048,0.009736888938479954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,256,0.014480888843536377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,3072,0.01761244402991401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,1536,0.006204444501135085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,256,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,2560,0.00942666663063897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,3072,0.007081777685218387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,10240,0.030511998467975195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,16384,0.08162044154273139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,3072,0.022675555613305833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,2560,0.006945778098371293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,1536,0.008642666869693333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,3072,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,65536,0.08100355333752103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,1024,0.00656355544924736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,4096,0.02108977735042572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,128,0.004934222333961063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,2560,0.006632888896597757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,5120,0.02643288837538825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,10240,0.023680888944202002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,6144,0.012402666939629449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,8192,0.019239111079110038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,8192,0.029967112673653498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,7168,0.02684533264901903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,65536,0.07211022244559394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,2048,0.004544888933499654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,4096,0.01737244427204132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,8192,0.015498666299713982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,16384,0.008338666624493068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,2048,0.010488889283604093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,5120,0.05239199929767185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,4096,0.03491111265288459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,16384,0.017799999978807237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,768,0.003216888962520493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,65536,0.15818933645884195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,4096,0.00906399968597624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,3584,0.01408266690042284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,4096,0.03393777873780992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,256,0.017854221993022494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,2560,0.022679110368092854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,256,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,3584,0.02809777855873108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,32,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,768,0.005280888742870755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,4096,0.007613333563009898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,10240,0.01860000027550591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,2048,0.013064888616402945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,128,0.013418667018413544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,768,0.002854222224818336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,12288,0.08126755555470784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,10240,0.012346666720178394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,5120,0.024295111497243244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,12288,0.02383111086156633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,12288,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,4096,0.025759999950726826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,3072,0.007231999602582719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,8192,0.016509332590632968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,8192,0.0218551109234492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,65536,0.11570933130052354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,1536,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,12288,0.06530577606625028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,7168,0.01015466699997584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,3584,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,768,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,2560,0.006592888798978593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,256,0.022974222898483276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,10240,0.022678222921159532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,768,0.0035217776894569397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,2048,0.007704888780911763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,1536,0.00518488883972168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,6144,0.04562577936384413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,128,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,64,0.0018684443914228016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,10240,0.022109333011839125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,12288,0.04336977667278714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,1024,0.013089777694808112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,6144,0.014356444279352823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,4096,0.011182222101423474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,6144,0.03880711065398322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,1536,0.01182755579551061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,3072,0.021807110971874658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,3584,0.036375999450683594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,128,0.0022257777551809945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,16384,0.23686220910814074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,128,0.0032479998966058097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,2560,0.006766222417354584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,65536,0.0846631129582723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,1536,0.016543111867374845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,1536,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,12288,0.03182844320933024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,65536,0.09052622318267822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,256,0.01896711190541585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,6144,0.013789333403110504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,3072,0.007266666326257918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,3584,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,65536,0.12921778361002603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,1536,0.09976355234781902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,16384,0.028431998358832464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,8192,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,10240,0.03185688787036472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,16384,0.021426666114065383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,2048,0.012228444218635559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,3584,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,10240,0.06265778011745878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,6144,0.02234399980968899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,1024,0.0120666664507654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,10240,0.011438222395049201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,1024,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,3072,0.06671910815768771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,16384,0.042488889561759106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,2048,0.03524444500605265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,1536,0.00655733338660664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,256,0.005643555687533484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,12288,0.0302471121152242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,768,0.015562666787041558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,768,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,6144,0.00674044465025266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,4096,0.11118222607506646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,2048,0.007629333270920648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,5120,0.15015021959940592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,256,0.014085332552591959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,8192,0.026554667287402686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,64,0.011016888750924004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,128,0.009747555686367882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,512,0.0025155554629034465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,10240,0.03576888971858554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,8192,0.07662310865190294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,65536,0.3092657725016276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,8192,0.010671999719407825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,65536,0.1761546664767795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,512,0.014473777678277759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,12288,0.01924977699915568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,64,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,12288,0.0362337761455112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,256,0.009380444056457942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,2048,0.0063182223174307085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,3072,0.007617777420414819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,768,0.0053342220683892565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,6144,0.012806221842765808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,6144,0.050903111696243286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,256,0.004895110925038655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,16384,0.024171556035677593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,3072,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,64,0.0831528902053833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,256,0.014056889547242058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,65536,0.10326221916410659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,3072,0.014102222190962898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,128,0.00489688871635331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,2048,0.022392888863881428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,64,0.0018835556175973679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,256,0.014127110441525778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,768,0.005513777749405966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,3072,0.013405332962671915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,256,0.016491555505328707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,5120,0.07430133554670545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,256,0.023393778337372675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,1536,0.020803555846214294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,4096,0.009139555195967356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,256,0.004888888862397936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,2560,0.01274400038851632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,512,0.008928889201747047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,768,0.011745777395036487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,65536,0.6941546863979764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,768,0.01716711123784383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,512,0.062372446060180664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,6144,0.011416888899273343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,512,0.009406222237481011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,16384,0.03727822171317206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,5120,0.02747111188040839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,65536,0.09693155686060588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,1024,0.008012444608741337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,16384,0.02437688906987508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,3072,0.047929777039421924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,8192,0.013393777940008374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,2560,0.00657155571712388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,1536,0.8761573367648654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,2560,0.008341333104504479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,3584,0.060198220941755504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,6144,0.015493333339691162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,32,0.0041999999019834726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,12288,0.03727466530270047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,4096,0.013421333498424955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,512,0.004853333450025982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,256,0.006280888699822956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,512,0.006990222467316522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,2560,0.009517333573765224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,4096,0.00795111142926746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,1024,0.009153777526484596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,3072,0.03487911158137851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,768,0.0031786666562159858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,32,0.001839111041691568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,3072,0.006928000185224745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,2560,0.010693333215183683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,2560,0.025624000363879736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,5120,0.01275022245115704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,128,0.01851288808716668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,128,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,768,0.015962666935390897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,3584,0.005938666562239329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,4096,0.031026668018764917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,2560,0.021316443880399067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,8192,0.04398933384153578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,128,0.01962311069170634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,10240,0.017701332767804463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,1024,0.007086222370465596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,512,0.016200888488027785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,8192,0.02366488840844896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,128,0.053114665879143604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,6144,0.016185777054892648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,8192,0.010771555205186209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,65536,0.10097777843475342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,3584,0.014691554837756686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,10240,0.15452888276841906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,2048,0.004556444370084339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,7168,0.04346844553947449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,5120,0.07379555702209473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,3072,0.00942844483587477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,512,0.0028248888750871024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,12288,0.013815999858909182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,3584,0.0075777777367168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,3072,0.08183111084832086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,12288,0.11843378014034694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,3584,0.024002666274706524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,65536,0.3081137869093153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,2048,0.011181332998805575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,12288,0.04954044355286492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,2048,0.006940444310506185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,256,0.01482488877243466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,5120,0.016515556308958266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,10240,0.0299253331290351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,64,0.0019120000716712738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,5120,0.03549066517088149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,1536,0.019080888893869188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,16384,0.05924266576766968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,512,0.004936000125275718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,10240,0.0633760028415256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,3584,0.007600888609886169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,12288,0.03353777858946059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,512,0.003847111016511917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,3072,0.006991110742092133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,128,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,12288,0.035678221119774714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,7168,0.016859556237856548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,10240,0.017881777551439073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,2560,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,512,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,256,0.01615111033121745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,2048,0.004252444538805219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,65536,0.07226488987604777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,256,0.007981333467695449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,4096,0.009700444837411245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,1536,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,64,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,512,0.026825777358478967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,1024,0.006038222048017714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,1024,0.007264888948864407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,16384,0.030174222257402208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,2560,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,1536,0.010761777559916178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,4096,0.008619555168681675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,16384,0.0780657794740465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,1024,0.01130844404300054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,65536,0.0878302190038893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,8192,0.010055111514197456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,16384,0.3057013352711995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,5120,0.03885777791341146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,512,0.003906666818592283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,4096,0.009487110707494948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,128,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,16384,0.021930666433440313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,768,0.0028195555011431375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,12288,0.018589332699775696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,64,0.0018026667336622875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,5120,0.03646577729119195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,1536,0.00600622221827507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,5120,0.01143555591503779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,6144,0.009319999979601966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,256,0.005615111026499007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,5120,0.014466666513019137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,12288,0.016544888416926067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,512,0.01348355578051673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,3072,0.07960622178183661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,2048,0.006232000059551663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,768,0.015145777000321282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,65536,0.09134399890899658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,2048,0.006523555351628198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,768,0.03642844491534763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,768,0.008676444490750631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,2560,0.006672888994216919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,768,0.01759288873937395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,128,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,12288,0.03477422065205044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,6144,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,5120,0.023006222314304773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,16384,0.03828622235192193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,512,0.013794666363133324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,1024,0.014485332700941296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,10240,0.014383999837769402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,3072,0.018596443865034316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,3072,0.14662755860222712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,5120,0.009297777381208207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,4096,0.011095999843544431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,10240,0.3439404434627957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,5120,0.033527112669414945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,6144,0.2323004404703776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,3584,0.024360888534122046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,65536,0.05779822005165947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,7168,0.021851556168662176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,1536,0.04420355624622769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,16384,0.055219554238849215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,10240,0.013730666703648038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,3584,0.009016888837019602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,3584,0.01055022246307797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,7168,0.03199911117553711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,1024,0.013759111364682516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,7168,0.00756888919406467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,128,0.015473778049151102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,8192,0.03481511274973551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,128,0.002143111079931259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,5120,0.019323555959595572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,4096,0.01682488951418135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,8192,0.02093688812520769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,6144,0.026100445124838088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,3584,0.016043555405404832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,1024,0.006919999917348226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,5120,0.00906133320596483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,6144,0.05602222018771701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,2560,0.022240888741281297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,10240,0.05658577548133003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,8192,0.04810933272043864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,3072,0.011432888607184092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,12288,0.04748533169428507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,2048,0.006619555668698416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,1024,0.03014844324853685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,10240,0.024688000480333965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,2048,0.019932443896929424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,8192,0.01034400032626258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,16384,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,32,0.001847111102607515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,2048,0.00998133338159985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,32,0.0057760001056724125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,7168,0.01199288914601008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,16384,0.005589333259397083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,128,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,8192,0.012797333300113678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,4096,0.013408888545301227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,768,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,512,0.021926222576035395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,65536,0.1610906653934055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,2560,0.01497066683239407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,6144,0.011618666350841522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,1536,0.005945777727497949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,1536,0.008657777474986183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,3072,0.01613066593805949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,2560,0.0069120000633928515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,768,0.01854222185081906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,512,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,8192,0.010672888822025724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,16384,0.022625777456495497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,16384,0.04034666551484002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,3072,0.030992888742023047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,16384,0.024396444360415142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,1024,0.005609777652555042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,65536,0.1596399943033854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,2048,0.028392887777752344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,1536,0.005980444451173146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,12288,0.06363466713163587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,65536,0.33255643314785427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,10240,0.020743111769358318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,6144,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,1024,0.006211555666393704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,768,0.005252444495757421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,16384,0.017504000001483493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,8192,0.06945066981845431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,512,0.020929776959949072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,3072,0.034930666287740074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,7168,0.23290933503044975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,8192,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,3584,0.007324444750944774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,1024,0.007295110987292395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,6144,0.04161599940723843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,1024,0.011741333537631564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,12288,0.028439998626708984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,128,0.013134222063753339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,10240,0.041482667128245033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,3072,0.020812445216708712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,2560,0.07641955878999498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,2048,0.006322666588756774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,16384,0.03189599845144484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,768,0.002850666642189026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,8192,0.013414222333166333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,1024,0.02106222179200914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,6144,0.009311999711725447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,1536,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,1024,0.012484444512261285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,3072,0.010448888772063784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,16384,0.016784888174798753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,3584,0.00904088881280687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,65536,0.05606844690110949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,128,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,256,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,5120,0.020449777444203694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,1024,0.00416355538699362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,256,0.014114666316244336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,1024,0.005622222191757626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,768,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,32,0.00702400008837382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,10240,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,3584,0.007247111035717859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,1024,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,4096,0.10143555535210504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,2560,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,32,0.0022071111533376905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,6144,1.832747565375434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,1024,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,2048,0.010032888915803697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,1536,0.004187555362780889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,10240,0.01792977750301361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,5120,0.015496888094478183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,1536,0.01921866668595208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,16384,0.03245155678855048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,65536,0.06413422028223674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,10240,0.051098667913013034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,16384,0.03749688797526889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,512,0.004873777843183941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,10240,0.040223111708958946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,3584,0.009309333231714036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,1536,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,16384,0.039340443081325956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,64,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,2560,0.008683555656009251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,1024,0.005456888841258154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,512,0.013065777719020844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,1536,0.0035217776894569397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,3584,0.007623111208279927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,256,0.01443555619981554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,6144,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,12288,0.035895999934938215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,512,0.00282577777074443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,2048,0.006970666348934174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,7168,0.523897753821479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,768,0.02125511070092519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,65536,0.010709332923094431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,3072,0.009015111459626092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,256,0.0038933331767717996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,768,0.05615378088421292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,12288,0.030947556098302204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,1024,0.03020355436537001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,3072,0.01382844481203291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,32,0.002171555534005165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,7168,0.015799111790127225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,512,0.01514844430817498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,64,0.003846222327815162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,10240,0.43971912066141766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,2048,0.07360000080532499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,4096,0.019496000475353666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,5120,0.025360888904995386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,1536,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,3072,0.0536097784837087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,1536,0.0032266666077905228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,256,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,32,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,8192,0.019934222102165222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,2048,0.004906666775544484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,8192,0.004253333227501975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,2560,0.10368088881174724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,3072,0.01330311099688212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,64,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,2048,0.08001689116160075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,6144,0.011071110765139261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,3072,0.015807999504937064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,8192,0.01851288808716668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,6144,0.09229511022567749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,7168,0.016559999850061204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,10240,0.008661333057615492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,6144,0.034966223769717746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,12288,0.018561777141359117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,3072,0.01683288812637329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,8192,0.012707555459605323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,4096,0.006372444331645966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,256,0.013739555246300168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,2560,0.020353777541054618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,512,0.0063235556913746735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,8192,0.1926186614566379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,768,0.017441777719391715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,2048,0.008999110923873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,1024,0.05076444480154249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,256,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,64,0.008310221963458592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,3584,0.010626667075686984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,768,0.016136000553766888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,3072,0.048135999176237315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,512,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,768,0.028944889704386394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,6144,0.0396364430586497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,10240,0.028970665401882593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,16384,0.0209840006298489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,128,0.005129777722888523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,65536,0.12596089310116237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,32,0.0038613333470291565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,128,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,12288,0.019596444235907663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,65536,0.01566844516330295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,8192,0.01036622209681405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,6144,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,512,0.003552888830502828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,1536,0.0035137778355015647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,65536,0.2765359878540039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,256,0.015815999772813585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,5120,0.009295110901196798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,512,0.013863999810483722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,768,0.1235840055677626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,2560,0.007281777759393056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,32,0.007972444097201029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,64,0.04010755485958523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,768,0.018259555101394653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,1024,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,6144,0.02622844444380866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,1024,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,5120,0.009947555760542551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,256,0.016128000285890367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,128,0.014520888527234396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,1024,0.007612444460391998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,64,0.001839111041691568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,16384,0.37647732098897296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,256,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,2048,0.005431111074156231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,12288,0.03580888774659898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,16384,0.025721778472264607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,3584,0.19539554913838705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,5120,0.00830666638082928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,2560,0.004867555366622077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,6144,0.014071111877759298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,128,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,2048,0.011707555916574268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,1536,0.013438222308953604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,1536,0.06783466868930392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,128,0.008651555412345463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,1024,0.011125333607196808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,1536,0.05348266826735603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,256,0.00286666676402092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,8192,0.019706666469573975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,5120,0.00889955543809467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,512,0.07149066527684529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,2560,0.013896889156765409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,3072,0.013436444103717804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,2048,0.04195911023351881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,256,0.0028204443968004654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,8192,0.020277332928445604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,5120,0.022636445032225713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,4096,0.006267555471923616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,1536,0.044222222434149853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,3072,0.012501333322789935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,10240,0.029318223396937054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,2560,0.01072177787621816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,768,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,2048,0.01868888901339637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,1024,0.004941333499219683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,256,0.016215110818545025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,1024,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,5120,0.01789777808719211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,2048,0.1325893269644843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,1024,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,256,0.014428445034556918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,3072,0.006956444846259222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,3072,0.022673777408070032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,256,0.01033866653839747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,16384,0.045392000012927584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,10240,0.25317690107557506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,256,0.0025066667132907445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,1536,0.027061333258946735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,256,0.005589333259397083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,10240,0.061963558197021484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,8192,1.328047964307997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,3072,0.00905511114332411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,16384,0.025045333637131587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,4096,0.009335111412737105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,5120,0.010064000056849586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,12288,0.03339733348952399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,512,0.02370044423474206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,5120,0.09766133626302083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,7168,0.03655644589000278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,1024,0.003569777641031477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,128,0.01767200065983666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,16384,0.3565688927968343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,32,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,32,0.0032053333189752367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,256,0.014174222946166992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,2560,0.017269333203633625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,3584,0.019668444991111755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,32,0.00589155571328269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,1536,0.01623288955953386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,7168,0.03601155678431193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,3072,0.025184889634450276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,10240,0.020346666375796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,768,0.004915555732117759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,8192,0.012424888710180918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,768,0.0038888889054457345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,65536,0.09391378031836616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,64,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,2560,0.004970666848950916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,65536,0.08847733338673909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,512,0.048801776435640126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,128,0.014112888111008538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,10240,0.017504000001483493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,10240,0.011713777979214987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,16384,0.12444267008039687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,5120,0.017519111434618633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,3072,0.007659555309348636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,8192,0.14089422755771214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,6144,0.020615110794703167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,32,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,1536,0.006685333533419504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,3072,0.006945778098371293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,128,0.011034666664070554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,1536,0.011694221860832639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,1536,0.008119111259778341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,512,0.05818400118086073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,65536,0.39864712291293675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,2048,0.013759999639458127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,16384,0.024803555674023096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,3072,0.007254222200976477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,3584,0.013805333111021252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,4096,0.10984622107611762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,8192,0.013833777772055732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,6144,0.016230222251680162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,5120,0.01826755536927117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,2048,0.021193777521451313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,5120,0.015842666228612263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,65536,0.12148355113135444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,32,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,768,0.002871111035346985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,3072,0.011693333586057028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,2560,0.0068897778789202375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,6144,0.009343999955389235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,3072,0.007637333538797166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,7168,0.026203556193245783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,3072,0.01886577738655938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,6144,0.016224000189039443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,4096,0.024736889534526404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,3584,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,6144,0.02126666737927331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,3584,0.03895911243226793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,1024,0.006931555353932911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,5120,0.09869600004620022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,65536,0.1824800040986803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,1536,0.004808889081080754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,5120,0.008645333349704742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,4096,0.032497776879204646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,7168,0.010023111270533668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,2048,0.006175111151403851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,64,0.003156444471743372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,2048,0.013772444592581855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,10240,0.03193688723776076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,5120,0.03375733229849074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,10240,0.026682666606373254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,128,0.005592000153329637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,2048,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,12288,0.03801244497299194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,10240,0.018617777360810172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,32,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,2560,0.02197333342499203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,65536,1.7578346464369032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,4096,0.03812088900142246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,768,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,3584,0.007538666327794393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,16384,0.14269599649641249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,65536,0.9168693754408094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,512,0.011382222175598145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,8192,0.018598222070270114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,4096,0.034860445393456355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,1024,0.00423555572827657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,128,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,768,0.0052986666560173035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,128,0.011767999993430244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,4096,0.14147555828094482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,2560,0.016152888536453247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,768,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,10240,0.021965333157115515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,1536,0.01107377807299296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,5120,0.1300062206056383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,32,0.006612444503439798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,4096,0.007245333658324347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,3584,0.014477334088749357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,6144,0.009224889179070791
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,1536,0.011599111060301462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,4096,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,65536,0.10263999965455796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,1536,0.197397338019477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,3584,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,8192,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,1024,0.013539555172125498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,16384,0.01720888912677765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,2560,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,3072,0.01037688884470198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,1536,0.014499555031458536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,3072,0.02756888833310869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,768,0.017909333109855652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,128,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,65536,0.07890222469965617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,2048,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,768,0.01770755483044518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,7168,0.02645955483118693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,4096,0.011752888560295105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,1024,0.005615111026499007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,65536,0.1836684412426419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,16384,0.021876444419225056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,2048,0.007654222349325816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,7168,0.06353155771891277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,10240,0.05061866508589851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,8192,0.012085333466529846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,2048,0.004587555511130227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,12288,0.022358222140206233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,128,0.014155555102560254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,512,0.023392889234754775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,1024,0.0031582222630580268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,128,0.013819555441538492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,1536,0.005964444329341252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,1024,0.005142222262091107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,1536,0.03583555420239767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,7168,0.013479111095269522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,1024,0.003600888782077365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,6144,0.009023110899660323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,1024,0.11971288257175022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,128,0.002124444478087955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,4096,0.01000533335738712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,1536,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,65536,0.5230097770690918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,16384,0.02988622254795498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,3072,0.27895911534627277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,5120,0.010613333019945355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,5120,0.06823199987411499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,768,0.004603555632962121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,2048,0.007637333538797166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,128,0.013432888521088494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,7168,0.01553066737122006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,12288,0.021332444416152105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,4096,0.06871555248896281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,1024,0.06436889039145575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,10240,0.030927111705144245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,3072,0.010713777608341642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,512,0.011423110961914062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,3584,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,768,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,256,0.02197066611713833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,65536,0.17456266615125868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,6144,0.009076444639099969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,1536,0.0038444445365005066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,2048,0.05635288688871595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,12288,0.023696889479955036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,1024,0.005935110979610019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,12288,0.013082666529549493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,2560,0.008963555925422246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,1536,0.012787555654843649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,3584,0.011386666860845355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,3584,0.015210666590266757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,16384,0.12919378280639648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,12288,0.01685599982738495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,768,0.017887110511461895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,1536,0.010818666881985135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,512,0.005277333160241445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,12288,0.26988622877332896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,32,0.006653333289755716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,10240,0.02100444502300686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,32,0.0018737777653667664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,65536,0.09535644451777141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,1024,0.006291555447710886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,128,0.00527555536892679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,5120,0.013400889105266996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,3072,0.016576889488432143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,3072,0.03545688920550876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,7168,0.01934400035275353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,4096,0.007642666498819987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,2560,0.010446222292052375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,7168,0.013785777820481194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,1536,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,8192,0.023013333479563396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,8192,0.03135644396146139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,2560,0.012519111235936483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,3584,0.04318844609790378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,128,0.008227555288208855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,3072,0.013437333206335703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,6144,0.009319110876984065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,10240,0.042690667841169566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,64,0.0018053334206342697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,3584,0.017567111386193168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,128,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,3584,0.007294221884674496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,8192,0.0960551102956136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,768,0.0038728887836138406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,4096,0.03530222177505493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,768,0.04496888981925117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,1536,0.005203555441564984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,32,0.005601777798599667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,3584,0.007592889169851939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,6144,0.009346666435400644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,32,0.00866844422287411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,7168,0.019293333093325298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,1024,0.0029315555261241067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,10240,0.6845982339647082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,7168,0.010701333483060202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,8192,0.016374222106403775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,1536,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,2560,0.023892444041040208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,1024,0.013064000341627332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,10240,0.01828533411026001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,4096,0.029072887367672388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,128,0.01623911162217458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,7168,0.3140328990088569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,12288,0.06071644359164768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,65536,0.1551191144519382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,12288,0.01899377836121453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,10240,0.02237333357334137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,256,0.014129777749379476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,1536,0.006866666591829724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,5120,0.13328888681199816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,1024,0.005877333382765452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,32,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,3072,0.006958222223652734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,1024,0.008995555341243744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,3584,0.008003555238246918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,768,0.0037271111375755737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,3584,0.02289955483542548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,6144,0.009187555147541894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,10240,0.017222222354676988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,7168,0.019824888971116807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,7168,0.010056888891590966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,1536,0.01686044368478987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,4096,0.02186222208870782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,12288,0.03259733319282532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,1024,0.007135111424658034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,65536,0.26231998867458767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,4096,0.007676444119877285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,3072,0.01688088807794783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,8192,0.1711519956588745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,7168,0.018751111295488145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,4096,0.01679377754529317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,4096,0.011809777882364062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,64,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,3584,0.014517333772447375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,768,0.008240888516108194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,32,0.00217333332531982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,3584,0.017696888910399545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,128,0.00554577757914861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,7168,0.05141866538259718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,3072,0.015229332778188916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,3584,0.007605333295133378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,12288,0.020285333196322124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,10240,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,3072,0.026591999663247004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,4096,0.04789244466357761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,6144,0.014751111467679342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,256,0.005081777771313985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,4096,0.007911110917727152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,5120,0.026904889278941687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,7168,0.013783999615245394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,2048,0.04548977812131246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,10240,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,4096,0.008264888491895463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,6144,0.009315555294354757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,128,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,128,0.005281777845488654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,65536,1.145897759331597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,5120,0.10621688763300578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,512,0.013456888496875763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,12288,0.014825777875052558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,256,0.005333333379692501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,3584,0.004867555366622077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,512,0.020594666401545208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,256,0.006503110958470239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,12288,0.022706665926509436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,65536,0.919099595811632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,3584,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,5120,0.01631733278433482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,7168,0.009645333720578088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,3072,0.00998133338159985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,4096,0.008363555702898238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,256,0.014226666755146451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,10240,0.004849777867396672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,7168,0.0141004439857271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,1536,0.020617778102556866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,3584,0.01661422186427646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,6144,0.014936000108718872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,16384,0.018862222631772358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,65536,0.011749332977665795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,65536,0.39303021960788304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,4096,0.016452444924248587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,5120,0.08729155858357747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,512,0.00798044436507755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,6144,0.013956444131003486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,6144,0.021311110920376245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,256,0.004910222358173794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,6144,0.03938577903641595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,7168,0.019556444552209642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,1536,0.006175111151403851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,1536,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,4096,0.011722666521867117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,512,0.0029084444459941653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,3584,0.006618666566080517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,5120,0.0046942221621672315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,10240,0.05065155691570706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,64,0.008991999758614434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,5120,0.017286222842004564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,64,0.006299555715587404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,128,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,6144,0.005938666562239329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,64,0.0021742222209771476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,256,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,1024,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,1024,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,3584,0.010608889162540436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,1536,0.007993777592976889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,2048,0.0047111109726958806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,1536,0.0041831110914548235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,1536,0.006131555471155379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,5120,0.008646222452322641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,128,0.004937777916590373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,6144,0.03295733200179206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,1536,0.012096888489193387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,512,0.01312444441848331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,4096,0.00793333351612091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,16384,0.01847644481394026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,5120,0.011743110915025076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,3584,0.0881955557399326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,512,0.028528001573350694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,3072,0.007279111279381647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,64,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,64,0.007311999797821045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,2560,0.0071928890215026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,3584,0.021387555532985266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,1536,0.006667555620272954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,7168,0.01311377767059538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,2560,0.006622222148709827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,3584,0.01518311103185018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,64,0.01424711114830441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,2048,0.011875555747085147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,256,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,7168,0.016206221448050607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,3584,0.01147822207874722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,16384,0.03343200021319919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,256,0.017235555582576327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,2048,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,3584,0.006923555499977536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,2048,0.018290667070282828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,7168,0.00983199973901113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,6144,0.025115556187099878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,8192,0.012103111379676394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,1536,0.005984888722499211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,2560,0.027540443672074214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,64,0.006636444479227066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,1536,0.006264888991912206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,10240,0.6238240136040581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,1024,0.0034808889031410217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,256,0.004933333231343163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,1536,0.05834933121999105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,65536,0.1400044494205051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,12288,0.02113155523935954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,6144,0.01369511087735494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,6144,0.11039466990364923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,1536,0.017445334129863314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,1536,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,65536,0.09842666652467515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,4096,0.018820444742838543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,32,0.0021866667601797315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,6144,0.017654221918847825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,2048,0.006647111227114995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,2560,0.020956445071432326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,5120,0.023390221926901076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,5120,0.02477600011560652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,5120,0.021553777986102637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,3584,0.02126400007141961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,16384,0.009879999690585667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,7168,0.39847644170125324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,8192,0.021983999345037673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,64,0.0025395556456512874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,2048,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,5120,0.08804088830947876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,4096,0.017673777209387887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,6144,0.014880888991885714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,64,0.009999111294746399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,7168,0.02699022160636054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,10240,0.034621334738201566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,64,0.0021884443445338146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,65536,0.16621955235799155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,2560,0.049999998675452344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,512,0.01681066718366411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,8192,0.2763333320617676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,5120,0.011058666639857821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,65536,0.10260622368918525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,5120,0.02409422232045068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,5120,0.0090506664580769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,12288,0.022375999225510493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,256,0.002165333264403873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,64,0.002526222210791376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,3072,0.013341333303186627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,3584,0.004857777721352047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,1536,0.031148443619410198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,1024,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,768,0.008274666965007782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,768,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,768,0.04850222335921394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,1024,0.003149333306484752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,4096,0.007616889145639207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,2560,0.15764089425404867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,6144,0.017017778423097398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,4096,0.022330666581789654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,3072,0.004583999928500918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,5120,0.008638222184446123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,5120,0.033956444925732084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,768,0.01239911135700014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,256,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,2560,0.02773777809407976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,5120,0.01809155609872606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,2048,0.004585777719815572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,256,0.0023680000255505242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,3072,0.005435555759403441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,5120,0.01314933349688848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,256,0.02774311105410258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,4096,0.015633778439627755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,2560,0.007347555624114142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,128,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,65536,0.06534488995869954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,1024,0.0035217776894569397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,16384,0.02884977724817064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,128,0.014440000057220459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,128,0.00544799988468488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,3072,0.02162044412559933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,65536,0.27923644913567436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,16384,0.015488889482286243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,65536,0.14084800084431967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,1536,0.019840000404251944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,16384,0.02632355524433984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,4096,0.013980444934633044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,512,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,1536,0.022683555881182354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,10240,3.9714168972439237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,3072,0.011410666836632622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,8192,0.07136178016662598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,32,0.009008888569143083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,8192,0.009334222310119206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,256,0.01538044379817115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,7168,0.010324444207880232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,256,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,8192,0.01752355529202355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,768,0.020272889071040683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,2048,0.031104889180925157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,8192,0.010781333678298526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,1536,0.006568888823191325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,256,0.01036888857682546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,1024,0.005974222388532426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,8192,0.04814222123887804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,1024,0.007334222396214803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,512,0.004899555610285865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,5120,0.021361778179804485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,4096,0.01036622209681405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,8192,0.039847112364239164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,512,0.01717333330048455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,65536,0.08695377906163533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,16384,0.06531466378106011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,3072,0.009143110778596666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,3072,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,2560,0.018907555275493197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,16384,0.02296000056796604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,16384,0.019818666908476088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,16384,0.07131110959582858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,3584,0.027169777287377253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,64,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,8192,0.013413333230548434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,64,0.006612444503439798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,2048,0.0069431112044387394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,6144,0.01718222267097897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,7168,0.04354666670163473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,3584,0.011053333679835001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,10240,0.0148053334818946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,4096,0.25433243645562065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,256,0.0025928889711697898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,64,0.006421333385838403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,3584,0.010484444598356882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,768,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,1024,0.0028168888141711554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,8192,0.05803110864427355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,1024,0.005429333282841577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,10240,0.0458568897512224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,4096,0.007272000114123027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,4096,0.012723555167516073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,65536,0.0780186653137207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,5120,0.15957244237263998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,3584,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,2560,0.006903111106819577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,512,0.00282222218811512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,5120,0.011323555476135679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,2560,0.008979555633332994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,1536,0.01680799987581041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,12288,0.014377777775128683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,2048,0.009820444716347588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,2560,0.01316088851955202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,768,0.01647822227742937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,64,0.006219555520349079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,7168,0.00870222184393141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,1536,0.0722915530204773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,128,0.003858666867017746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,8192,0.04087022278043959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,128,0.014155555102560254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,16384,0.06481155421998766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,1536,0.04601777924431694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,128,0.011756444142924415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,12288,0.15245778030819362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,3584,0.007280000381999546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,2560,0.00665244460105896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,128,0.006591111007663939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,1024,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,1536,0.014455111490355598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,12288,0.0199644449684355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,64,0.007307555940416124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,5120,0.00847111145655314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,32,0.03277600142690871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,3072,0.014205333259370593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,768,0.030378666188981798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,65536,1.3235342237684462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,4096,0.02899822261598375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,1536,0.014479110638300577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,2048,0.006692444284756978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,6144,0.06306578053368463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,65536,0.11979556083679199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,256,0.004934222333961063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,1024,0.009012444151772393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,768,0.0029120000286234748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,8192,0.08057777749167548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,1536,0.003536000019974179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,768,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,128,0.013753777576817406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,8192,0.019165333774354722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,2048,0.006454222318198945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,1024,0.014631999863518609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,512,0.0058764442801475525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,1024,0.003128000017669466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,2560,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,512,0.014155555102560254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,3584,0.06875733534495036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,16384,0.029987557066811457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,128,0.016757332616382174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,1024,0.011397333608733283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,4096,0.011412444214026133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,12288,0.041568888558281794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,64,0.011408888631396823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,8192,0.013791999883121915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,1536,0.0038720000949170855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,32,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,512,0.013550221920013428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,65536,0.04514222343762716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,768,0.019315555691719055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,1536,0.010460444622569613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,4096,0.04107022285461426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,256,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,65536,0.07875466346740723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,1536,0.0075786668393347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,2048,0.10833155446582371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,6144,0.013146667016877068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,65536,0.12753244241078696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,16384,0.0339626669883728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,3072,0.010456889039940305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,128,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,8192,0.044941332605150014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,1024,0.011018666956159802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,128,0.020976889464590285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,16384,0.1418559948603312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,32,0.006355555521117316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,6144,0.06650844547483656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,2048,0.027826666831970215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,1024,0.00610133343272739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,1024,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,512,0.0028195555011431375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,512,0.0032231110251612137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,2048,0.012162666353914471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,5120,0.012868444124857584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,10240,0.028888887829250757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,65536,0.07556711302863227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,8192,0.02366222275627984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,10240,0.023453333311610754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,32,0.0018888889915413326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,3584,0.02273155583275689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,1024,0.003206222214632564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,2560,0.057348443402184375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,6144,0.041815999481413096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,128,0.005930666708283954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,4096,0.15456977155473497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,256,0.006706666615274217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,2048,0.006279110908508301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,128,0.004939555707905027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,32,0.013794666363133324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,10240,0.018222222725550335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,256,0.004965333475006951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,5120,0.012358222570684222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,1024,0.044814222388797335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,5120,0.02784088916248745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,8192,0.0192586663696501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,65536,0.06583288643095228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,2048,0.008341333104504479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,8192,0.005589333259397083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,32,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,3584,0.00999022192425198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,2560,0.01644000079896715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,512,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,512,0.007697777615653143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,2560,0.05572622352176242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,2560,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,7168,0.02364088926050398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,4096,0.015202666322390238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,10240,0.017299556069903903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,8192,0.01423377792040507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,256,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,1536,0.06779466734992133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,1024,0.01923822197649214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,2560,0.019831111033757527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,8192,0.04100711146990458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,2048,0.027706666125191584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,65536,0.16052977244059244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,4096,0.020440888073709276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,2560,0.06950222121344672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,12288,0.010246222217877706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,1024,0.011389333340856763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,512,0.014517333772447375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,2048,0.004206222378545337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,2560,0.04011733333269755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,7168,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,768,0.01818933255142636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,16384,0.012574222352769641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,5120,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,12288,0.022358222140206233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,128,0.03539289037386576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,128,0.004918222212129169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,2560,0.02195466723706987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,2560,0.006640000061856375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,3584,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,4096,0.03325600094265408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,1024,0.012740444805887012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,128,0.002565333412753211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,2560,0.004584889031118817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,10240,0.013919111755159167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,256,0.005142222262091107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,5120,0.012113778127564324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,6144,0.011734222372372946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,256,0.11178844504886204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,1536,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,8192,0.011771555576059552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,1024,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,6144,0.013356444736321768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,3584,0.0314355558819241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,256,0.021419554948806763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,2048,0.008630221916569604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,3072,0.00835111074977451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,3072,0.007280000381999546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,512,0.0032044444233179092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,4096,0.007483555210961237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,2560,0.028169777658250596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,7168,0.5546711285909017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,7168,0.01310488912794325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,4096,0.02161955502298143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,16384,0.26964444584316677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,2048,0.014484445254007975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,512,0.0034933334423436057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,2560,0.008003555238246918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,2048,0.006232888748248418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,12288,0.3496844503614638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,10240,0.04240622123082479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,12288,0.13082488377888998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,32,0.005992888990375731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,12288,0.013184000220563678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,512,0.006656888872385025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,7168,0.060989333523644336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,10240,0.012136889000733694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,128,0.0025902222841978073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,2560,0.012141333685980903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,2560,0.013479111095269522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,3072,0.020670221911536325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,1536,0.0083137775460879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,3584,0.009321778184837764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,256,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,2560,0.023043556345833674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,512,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,1536,0.006301333506902059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,12288,0.07404800256093343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,3072,0.008649778034951951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,4096,0.022075555390781824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,768,0.014520888527234396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,512,0.005239999956554837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,256,0.01108977778090371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,256,0.014106666048367819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,16384,0.02914666798379686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,3584,0.02309777836004893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,64,0.006600889066855113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,12288,0.06710488928688897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,256,0.11479289001888698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,768,0.012086222569147745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,1536,0.004219555606444676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,65536,1.9915794796413844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,4096,0.04939644535382589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,10240,0.021293333835071985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,8192,0.01682400041156345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,7168,0.11902399857838948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,8192,0.0496711101796892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,6144,0.33636977937486434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,12288,2.4180480109320746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,16384,0.016872000363137987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,6144,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,128,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,12288,0.039134220944510564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,6144,0.016887111796273124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,256,0.013803555733627744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,2048,0.009219555391205682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,2560,0.028837333122889202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,65536,0.15988711516062418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,64,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,16384,0.045315557056003146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,3072,0.007304888632562425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,1536,0.010018666585286459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,3072,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,1024,0.0058702222175068325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,16384,0.02514311174551646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,16384,0.021292444732454088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,10240,0.015789333317014907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,256,0.01036177741156684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,1024,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,65536,0.07880533403820462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,1536,0.006893333461549547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,10240,0.01578044394652049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,65536,0.07518133189943102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,512,0.020283554991086323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,3072,0.017210667332013447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,65536,0.4787751303778754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,4096,0.008559111091825698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,512,0.032976001501083374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,10240,0.006609777609507243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,6144,0.011710222396585675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,64,0.0021084443562560612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,2048,0.010816000401973724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,12288,0.09824355443318684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,10240,0.07022933165232341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,12288,0.08578755458196004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,3072,0.01849600010448032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,1536,0.009344889058007134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,12288,0.014114666316244336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,512,0.013415999710559845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,768,0.0033946666452619764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,64,0.01685244507259793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,65536,0.0780986679924859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,6144,0.02421599957678053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,3584,0.008013333711359236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,2048,0.01685244507259793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,5120,0.02130577796035343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,4096,0.03177244464556376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,64,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,2560,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,5120,0.014461333552996317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,6144,0.014601776997248331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,4096,0.00869511150651508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,65536,0.054811557133992515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,1536,0.010836444795131683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,10240,0.017157332764731515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,1024,0.004000888930426704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,8192,0.011745777395036487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,4096,0.00868000007337994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,3072,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,512,0.006272000157170826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,12288,0.02332177758216858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,10240,0.004854222138722737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,128,0.0038693332009845306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,32,0.0028808888875775864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,4096,0.0079297779334916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,10240,0.03294666608174642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,1536,0.012776888906955719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,65536,0.18480177720387778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,4096,0.01682222220632765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,4096,0.011951111257076263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,8192,0.014444443914625378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,65536,0.6839777628580729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,3072,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,128,0.00250133333934678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,2560,0.021906667285495337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,512,0.0031431110368834603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,65536,0.08559644222259521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,8192,0.013983110586802164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,3072,0.007274666594134436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,256,0.005950222412745158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,2560,0.0203164451652103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,16384,0.023961777488390606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,5120,0.017117333081033494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,16384,0.029106666644414265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,5120,0.02290222214327918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,8192,0.03875822159979079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,128,0.0022195554855797025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,3584,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,4096,0.008314666648705801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,3584,0.004931555440028508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,2560,0.019574221637513902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,2048,0.013818666338920593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,64,0.0302817788388994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,8192,0.01072266697883606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,256,0.01146399974822998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,128,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,32,0.002536000063021978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,3072,0.017288888494173687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,4096,0.007634667058785756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,16384,0.5296693378024632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,768,0.00544177782204416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,5120,0.017250667015711468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,5120,0.004182222402758068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,16384,0.017924444542990792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,1536,0.017877333694034152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,512,0.002530666689078013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,256,0.002452444492114915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,7168,0.019282667173279654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,1024,0.009307555854320526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,5120,0.4958817693922255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,256,0.002518222149875429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,128,0.014114666316244336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,6144,0.04615199897024366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,10240,0.018221333622932434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,12288,0.02275644408331977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,512,0.005490666462315454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,1024,0.01590755581855774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,4096,0.008591111335489485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,2048,0.013389333254761167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,16384,0.042913777960671314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,1536,0.018201778332392376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,6144,0.01854044364558326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,8192,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,64,0.0021111110432280437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,16384,0.023013333479563396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,1024,0.06510311365127563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,256,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,1536,0.00795377790927887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,5120,0.017503110898865592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,65536,0.05367999937799242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,1024,0.006627555522653792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,12288,0.03483555714289347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,2048,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,4096,0.08339111010233562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,64,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,12288,0.021353777911927965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,5120,0.1377448903189765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,32,0.003236444460021125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,3072,0.011072888970375061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,5120,0.01276266657643848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,5120,0.00739822205570009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,4096,0.00870400004916721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,32,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,3584,0.07898222075568305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,4096,0.017701332767804463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,256,0.007285333342022366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,16384,0.02126400007141961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,2048,0.013154666456911298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,2560,0.007393777370452881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,5120,0.028710222906536523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,16384,0.3647768762376573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,1024,0.006535111202134027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,8192,0.015234667393896313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,6144,0.043488889932632446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,3072,0.010062221851613786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,512,0.005912888795137405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,256,0.01647111111217075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,64,0.007620444728268518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,256,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,4096,0.00870311094654931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,3584,0.011316444310877057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,128,0.011362666885058085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,64,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,768,0.005919111271699269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,2048,0.0094035557574696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,8192,0.020251555575264823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,12288,0.013777777552604675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,768,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,256,0.002471111093958219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,10240,0.02299022177855174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,1024,0.00573955559068256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,32,0.00564444437623024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,1536,0.005620444400442972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,2048,0.00452888881166776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,12288,0.018838221828142803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,1024,0.004856889032655292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,256,0.005608888963858287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,3072,0.007310222420427535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,512,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,768,0.006228444476922353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,10240,0.02749155627356635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,128,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,16384,0.02947288751602173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,768,0.005960888746711943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,12288,0.016484444340070088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,768,0.018535999788178336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,3072,0.05923555294672648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,2560,0.009544889132181803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,3072,0.007654222349325816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,256,0.005239999956554837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,2560,0.01645511057641771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,1536,0.00832622249921163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,1536,0.01110311100880305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,3584,0.026390223039521113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,1536,0.007674666742483775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,1536,0.013089777694808112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,6144,0.009798222117953831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,12288,0.6736453374226888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,768,0.002999111182159848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,128,0.0021173333128293357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,12288,0.010071111222108206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,65536,0.08928000264697605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,7168,0.17873777283562553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,8192,0.11345511012607151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,3072,0.00869333330127928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,12288,0.013456000222100152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,16384,0.04301066531075371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,32,0.007596444752481248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,65536,0.05526044302516513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,5120,0.013400889105266996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,32,0.0018071112119489247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,8192,0.166049771838718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,3584,0.00795111142926746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,12288,0.019670221540662978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,1536,0.013174222575293647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,5120,0.03373333480623033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,1536,0.00794400026400884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,6144,0.24159110916985405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,16384,0.04296355446179708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,32,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,6144,0.02324177821477254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,1536,0.003522666792074839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,2048,0.017854221993022494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,1024,0.014135110709402295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,12288,0.08368533187442356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,6144,0.16760978433820936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,2560,0.005950222412745158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,3584,0.020625778370433383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,7168,0.029878222280078467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,3072,0.017151999804708693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,7168,0.013804444836245643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,128,0.005249777601824866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,32,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,128,0.00795377790927887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,3072,0.004859555512666702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,2560,0.012819555898507437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,3584,0.012426666915416718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,64,0.0018133332745896445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,4096,0.015447111593352424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,7168,0.009660444325870937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,10240,0.012089778151777057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,3072,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,8192,0.01624000072479248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,10240,0.012042666474978128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,2560,0.021222222182485793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,16384,0.23242934544881186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,6144,0.011412444214026133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,2048,0.022294221652878657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,64,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,8192,0.016709332664807636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,768,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,5120,0.019282667173279654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,1536,0.007617777420414819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,1024,0.022498667240142822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,2560,0.04012889001104567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,3584,0.0057502223385704895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,256,0.014374223020341663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,6144,0.03166933192147149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,16384,0.03273866573969523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,3584,0.0425217780801985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,3584,0.015912888778580558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,8192,0.010749333434634738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,1024,0.0029022221763928733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,8192,0.03091555502679613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,2048,0.01313777764638265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,64,0.0028524444335036804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,65536,0.017847110827763874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,1536,0.015110222829712762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,6144,0.02791555391417609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,64,0.008284444610277811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,3584,0.017824000782436795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,256,0.004912888838185204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,65536,0.10453599691390991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,512,0.005886222339338726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,64,0.001995555642578337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,1024,0.0535715553495619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,128,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,32,0.005592000153329637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,512,0.005534222142563925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,32,0.01071111112833023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,512,0.010672888822025724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,65536,3.5978143480088978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,128,0.0086666668454806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,10240,0.023453333311610754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,10240,0.007629333270920648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,3584,0.09458666377597386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,4096,0.0673324465751648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,2560,0.08319733540217082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,3584,0.059791107972462974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,2048,0.015501333607567681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,65536,0.9507012897067599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,256,0.004551110996140374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,128,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,2048,0.006613333192136552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,6144,0.026535110341178045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,6144,0.00909511082702213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,3072,0.019641778535313077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,1536,0.008657777474986183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,2560,0.015840000576443143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,3072,0.015844444433848064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,3072,0.01074399964676963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,12288,0.01572088897228241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,1536,0.005961777849329843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,128,0.004941333499219683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,6144,0.009193778038024902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,32,0.007671111159854465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,12288,0.018912000788582694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,3584,0.014627556006113688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,2560,0.0069324444565508105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,7168,0.024490666058328416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,128,0.014821334017647637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,8192,0.007599111232492659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,12288,0.016414221790101793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,32,0.003825777934657203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,128,0.002536000063021978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,8192,0.011077333655622272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,128,0.014471999473041959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,16384,0.048771556880739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,8192,0.02276355524857839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,8192,0.08606044451395671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,6144,0.009325332939624786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,128,0.013812444276279874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,1024,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,7168,0.012040889097584618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,256,0.002830222249031067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,10240,0.01237511138121287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,3072,0.010759111079904767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,16384,0.07894666989644368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,4096,0.04821955495410495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,12288,0.058450665738847524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,768,0.007918222082985772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,256,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,2048,0.015460444821251763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,8192,0.02184355590078566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,5120,0.011704000333944956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,768,0.00563822231358952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,1536,0.005912000106440649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,12288,0.04057244459788004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,7168,0.03035555614365472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,256,0.018204443984561496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,2560,0.009429333110650381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,3072,0.03474488854408264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,2560,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,16384,0.0232622226079305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,7168,0.11165600352817112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,128,0.0053102220926019884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,4096,0.016855110724767048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,16384,0.07836444510353936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,6144,0.011952000359694162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,1024,0.07837600178188749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,8192,0.013459555804729462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,4096,0.08029155598746406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,16384,0.021485333641370136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,256,0.016859556237856548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,12288,0.057736886872185596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,768,0.014555555250909595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,7168,0.03815111186769273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,8192,0.02955555584695604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,512,0.007107555866241455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,2560,0.013838222457302941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,8192,0.024039111203617517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,10240,0.012676444318559436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,256,0.015139554937680563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,3584,0.031089779403474595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,512,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,10240,0.024523556232452393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,128,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,512,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,256,0.018942221999168396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,768,0.0032257777121331957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,4096,0.015559110376569959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,1024,0.007292444507280986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,3584,0.01945333348380195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,2560,0.006904888898134232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,512,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,768,0.003286222202910317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,1536,0.006285333385070165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,16384,0.03526399864090814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,3584,0.004935111022657818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,3072,0.009025777379671732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,1536,0.006220444622966979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,128,0.004967999955018361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,8192,0.05806489123238457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,65536,0.07566044727961223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,8192,0.014151111245155334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,16384,0.05074577861362033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,6144,0.03323288758595785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,7168,0.028451555305057104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,1536,0.014527999692493014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,768,0.017267554998397827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,3072,0.007240888973077138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,10240,0.01757333344883389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,3584,0.0406640006436242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,2560,0.042295111550225147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,10240,0.2320480081770155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,2560,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,6144,0.013584000368913015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,2560,0.007710222568776872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,2560,0.009092444347010719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,512,0.012412444584899478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,7168,0.014529777897728814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,2048,0.00720177756415473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,768,0.005169777820507686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,7168,0.09815733300315009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,128,0.013785777820481194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,10240,0.035114665826161705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,3584,0.00793422261873881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,256,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,3584,0.023768888579474554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,7168,0.14703910880618626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,12288,0.03011911114056905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,3072,0.006934222247865465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,3584,0.012090666426552666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,6144,0.012294222083356647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,1024,0.012072888513406118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,3584,0.06141777833302816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,10240,0.025051555699772302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,128,0.007575111256705389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,8192,0.01721955504682329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,7168,0.23268177774217394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,512,0.016545777519543965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,768,0.005296000176005893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,512,0.002480888946188821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,8192,0.007602666815121968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,512,0.005304000029961268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,2560,0.09858577781253391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,4096,0.20468444294399687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,12288,0.04462488823466831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,256,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,2048,0.07669066720538668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,16384,0.020865778128306072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,512,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,768,0.005118222286303838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,256,0.01443199978934394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,4096,0.009701333112186855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,65536,0.2525662316216363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,2560,0.010394666757848527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,1024,0.09167644712660049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,3072,0.013844444519943662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,1024,0.0233324451578988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,1024,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,2560,0.004856889032655292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,2048,0.006270222365856171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,65536,0.216538667678833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,512,0.004979555391603046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,7168,0.03736266824934218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,16384,0.3575102223290338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,16384,0.9366444481743706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,768,0.016120000018013846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,5120,0.04342488778962029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,12288,0.09630311197704738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,768,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,6144,0.02757688860098521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,768,0.12068000104692246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,3584,0.026144888665941026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,2560,0.00812355594502555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,256,0.014841778410805596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,1024,0.006018666757477655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,768,0.020922667450375024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,3072,0.015473778049151102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,16384,0.1390897830327352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,32,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,16384,0.04871822065777249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,6144,0.009023110899660323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,2560,0.01583377851380242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,16384,0.02346666653951009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,7168,0.010440000229411654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,64,0.006242666807439592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,5120,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,16384,0.4569137891133626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,7168,0.07416711250940959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,8192,0.06392977635065715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,128,0.006953777538405524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,32,0.010648888846238455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,64,0.010497777826256223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,4096,0.037893331713146634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,128,0.02899644441074795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,65536,0.0762631098429362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,12288,0.022640888889630634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,7168,0.02734044525358412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,768,0.0042168887125121225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,4096,0.033570667107899986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,16384,0.026084444589085046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,256,0.0024764444679021835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,8192,0.01643733349111345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,3072,0.009335111412737105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,128,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,3584,0.007329777710967594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,2560,0.009668444593747457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,8192,0.015819556183285184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,65536,0.0984755555788676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,128,0.013960889644092984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,3584,0.0094035557574696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,512,0.009009777671760982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,3584,0.027058665951093037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,32,0.08613955312305027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,12288,0.06352711386150785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,1536,0.023363555471102398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,10240,0.030357334348890517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,16384,0.4948479864332411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,65536,0.04558222161398994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,2560,0.023295111126369897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,6144,0.009731555150614845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,6144,0.028590222199757893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,512,0.004518222063779831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,2048,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,65536,0.09220621983210246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,512,0.016896888613700867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,6144,0.12030488914913601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,6144,0.022896000080638464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,6144,0.012488889197508493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,5120,0.019882667395803664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,12288,0.07376266850365533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,128,0.006248000181383557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,6144,0.03661333190070258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,16384,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,6144,0.009726222190592024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,4096,0.020587555236286588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,3584,0.012727999852763282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,6144,0.009039999710188972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,1024,0.03156800071398417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,5120,0.03217511044608222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,5120,0.02639377779430813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,32,0.007610666255156199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,10240,0.016494222813182406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,64,0.0042453333735466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,5120,0.00666844430896971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,10240,0.016511999898486666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,8192,0.02536444365978241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,3072,0.014789332946141561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,256,0.0026666666898462507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,128,0.006242666807439592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,4096,0.009024889104896123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,3584,0.008309333688682979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,3584,0.020623111062579684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,5120,0.2907608879937066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,1536,0.0916408896446228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,3072,0.023430221610599097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,6144,0.01274400038851632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,1024,0.0031893334041039148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,12288,0.0544177790482839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,128,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,1024,0.005625777774386936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,2048,0.011064888702498542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,16384,0.22155733903249106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,65536,0.07587377892600165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,7168,0.011069333387745751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,2560,0.039748443497551814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,32,0.006263999889294307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,16384,0.0324888891643948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,10240,0.012051555845472546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,8192,0.3474115530649821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,256,0.005033777819739448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,64,0.0029395555870400537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,10240,0.0162426663769616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,10240,0.027270222703615826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,2048,0.009073778159088558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,8192,0.14134844144185385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,1024,0.0073520003093613525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,16384,0.02201066745652093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,2048,0.01756977703836229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,7168,0.013368888861603208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,8192,0.009708444277445475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,2048,0.018194667167133756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,16384,0.04353777898682488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,3584,0.03871555460823907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,12288,0.01574488812022739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,2048,0.019289778338538278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,12288,0.1041075587272644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,3584,0.015497777197096082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,3584,0.08766400151782566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,128,0.013415111435784234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,2560,0.006928000185224745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,256,0.00216977774269051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,4096,0.015468445089128284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,32,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,1536,0.04899200134807163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,16384,0.0210506667693456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,768,0.01792622274822659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,2048,0.016157334049542744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,2560,0.008616000413894653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,5120,0.005384888913896348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,5120,0.010061333576838175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,768,0.028181334336598713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,128,0.014104000396198697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,6144,0.021652445197105408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,7168,0.02295555505487654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,768,0.01685244507259793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,7168,0.026935999592145283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,32,0.051855110459857516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,1024,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,5120,0.03281955586539374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,7168,0.006928000185224745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,10240,0.008640888664457533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,1024,0.00452622233165635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,2560,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,6144,0.01441422270403968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,16384,0.030262221892674763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,1536,0.02380888826317257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,65536,0.23230489095052084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,10240,0.02785244584083557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,6144,0.013071999781661563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,8192,0.01034577770365609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,128,0.005048888838953442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,3584,0.053935110569000244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,256,0.006305777778228124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,1024,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,1536,0.028215110301971436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,768,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,128,0.020943111843532983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,7168,0.01236977759334776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,6144,0.016916443904240925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,16384,0.017087999317381117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,2560,0.009008000294367472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,256,0.014454222387737699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,12288,0.01514400045077006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,3584,0.008756444685988957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,5120,0.00869511150651508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,65536,0.20083644655015734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,12288,0.05950488646825155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,128,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,128,0.004577777865860197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,7168,0.004912888838185204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,128,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,5120,0.0240000006225374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,10240,0.026708443959554035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,2048,0.015695111619101632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,768,0.016498666670587327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,256,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,256,0.006248000181383557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,16384,0.02972888946533203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,3072,0.02520977788501316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,3072,0.021685333715544805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,4096,0.020522667302025687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,12288,0.37931733661227757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,4096,0.023984000086784363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,1536,0.05195555422041151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,512,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,512,0.005576000031497743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,3584,0.06638755400975545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,512,0.03349422083960639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,8192,0.02490844494766659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,2048,0.02686311139000787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,128,0.007284444239404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,3072,0.18393154939015707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,768,0.004581333448489507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,2048,0.013110222087966071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,3072,0.00423555572827657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,128,0.0020622221959961788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,512,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,512,0.006337777607970768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,1024,0.014125333891974555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,2560,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,3584,0.010876444478829702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,6144,0.03243288728925917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,6144,0.030435555511050757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,16384,0.020970667401949566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,16384,0.05905066596137153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,8192,0.039649777942233615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,128,0.006251555350091722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,128,0.010703110860453712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,1024,0.007307555940416124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,3584,0.008030222521887885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,2560,0.006637333167923822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,12288,0.014440889159838358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,5120,0.026865777042176988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,3072,0.01036177741156684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,8192,0.010393777655230628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,16384,0.3532080120510525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,256,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,3584,0.032241778241263494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,7168,0.026072000463803608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,65536,0.12366044521331787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,12288,0.029273778200149536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,10240,0.017857778403494094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,2048,0.03121244576242235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,2048,0.01777866648303138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,6144,0.03246488836076524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,5120,0.03864799936612447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,16384,0.026159110996458266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,32,0.005917333480384614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,2048,0.006264888991912206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,768,0.029048889875411987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,10240,0.014821334017647637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,3072,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,1024,0.006586666736337874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,8192,0.042881776889165245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,5120,0.01076533314254549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,4096,0.0075902218619982404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,16384,0.1011155578825209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,2560,0.00721866637468338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,768,0.008962666822804345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,64,0.006232888748248418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,16384,0.012668444050682915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,2048,0.007032000356250339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,2048,0.008626666333940294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,2048,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,128,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,768,0.002934222213096089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,65536,0.06407644351323445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,5120,0.1071573363410102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,12288,0.017635555730925668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,12288,0.013073777986897362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,128,0.004559999952713649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,128,0.02573066618707445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,6144,0.23594400617811415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,512,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,3584,0.021018667353524104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,1024,0.007618666523032718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,5120,0.017676444517241586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,12288,0.027458666099442378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,65536,0.22011200586954752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,128,0.002141333288616604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,1024,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,3072,0.006632888896597757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,3584,0.00796088907453749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,7168,0.004960000101062986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,1536,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,128,0.006257777826653586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,2560,0.007266666326257918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,1024,0.01074222226937612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,5120,0.00889066689544254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,10240,0.012108444339699216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,512,0.003565333369705412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,10240,0.0122079998254776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,1536,0.018738667170206707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,3584,0.016750221451123554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,1024,0.003112000102798144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,12288,0.028047998746236164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,2560,0.013745778136783175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,128,0.013975999421543546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,2048,0.06905600097444323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,1024,0.012440000143316058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,1536,0.00830577810605367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,12288,0.1213466723759969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,3072,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,3584,0.01019377758105596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,4096,0.012797333300113678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,7168,0.02271644439962175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,128,0.008416000339719985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,6144,0.013436444103717804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,768,0.02418400016095903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,7168,0.013012444807423485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,2560,0.00664444433318244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,3584,0.007639110916190677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,64,0.003134222287270758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,256,0.004967999955018361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,3072,0.005305777821275923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,16384,0.07893066936069064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,10240,3.5294187333848743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,64,0.009176000124878353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,16384,0.41404088338216144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,1024,0.016143111719025504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,2560,0.028206222587161597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,512,0.010017777482668558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,6144,0.1155173381169637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,256,0.004895110925038655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,128,0.013772444592581855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,10240,0.03035911255412632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,10240,0.02548177705870734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,2048,0.0041777777175108595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,768,0.006228444476922353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,10240,0.014491554763582019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,1536,0.006619555668698416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,8192,0.010406222608354356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,3584,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,10240,0.38071288002861875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,256,0.002187555655837059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,12288,0.022446221775478784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,768,0.009135111338562435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,4096,0.011402666568756104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,65536,0.09742577870686848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,6144,0.03372800019052293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,7168,0.004884444591071871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,6144,0.011700444751315646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,4096,0.023592000206311543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,128,0.012663111090660095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,256,0.011543110840850405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,1536,0.023585778143670823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,1536,0.00629333323902554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,2048,0.008656889200210571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,32,0.0021137777302000257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,65536,0.1522159973780314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,1536,0.018239110708236694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,1536,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,768,0.005525333185990651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,4096,0.027141332626342773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,1024,0.0058497778243488735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,2560,0.007038222418891058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,2560,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,256,0.002521777732504739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,1536,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,1536,0.03765599926312765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,10240,0.016509332590632968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,1536,0.005867555737495422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,5120,0.008606221940782335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,3072,0.019511999355422128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,8192,0.037252446015675865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,16384,0.07527022229300605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,1024,0.01716711123784383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,512,0.0042257776690853965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,512,0.003083555648724238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,512,0.014819555812411837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,64,0.0018462222069501877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,12288,0.01959911154376136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,3584,0.00756444450881746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,3584,0.01590844492117564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,65536,0.07936266395780775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,1536,0.006148444281684027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,16384,0.023032888770103455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,64,0.0018515555808941524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,2048,0.024392000503010217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,4096,0.02379111117786831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,5120,0.008630221916569604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,4096,0.015544889701737298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,256,0.0032435556252797446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,3584,0.025783111651738484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,256,0.014487110906177096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,1024,0.018950222267044913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,4096,0.006958222223652734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,2560,0.006267555471923616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,64,0.006280888699822956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,16384,0.02426133387618595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,8192,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,65536,0.07621155844794379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,1536,0.03756533397568597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,1536,0.006069333189063602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,1536,0.0035253332720862497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,3584,0.016520000166363187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,32,0.005943111247486538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,6144,0.027961777316199407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,4096,0.032425777779685125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,128,0.0051893335249688895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,7168,0.027887112564510767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,1536,0.07912088765038384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,16384,0.020664888951513503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,64,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,2048,0.008529778156015608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,65536,1.9334017435709636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,5120,0.025807110799683466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,2560,0.006994666324721442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,2560,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,10240,0.013082666529549493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,10240,0.47821866141425234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,2048,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,4096,0.030552887254291113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,8192,0.04055555661519369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,512,0.003403555601835251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,4096,0.007970666719807519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,8192,0.02645066711637709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,1024,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,1024,0.00943911075592041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,7168,0.03657244311438666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,2560,0.033210668298933245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,16384,0.07861955298317803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,32,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,1024,0.0031822222388452957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,8192,0.018212444252438016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,3584,0.007659555309348636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,8192,0.01499288943078783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,12288,0.05433600147565206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,512,0.006587555425034628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,2048,0.012679110798570844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,10240,0.018212444252438016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,3584,0.02347466680738661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,1024,0.012841777669058906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,1024,0.007230222225189209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,768,0.005428444594144821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,512,0.0028488888508743713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,2560,0.012468444804350534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,8192,0.024525332782003615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,768,0.01587377819750044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,2560,0.028348445892333984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,6144,0.019898666275872123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,7168,0.032587554719712995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,1024,0.01587466730011834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,1024,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,768,0.02402488887310028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,256,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,65536,0.06268088685141669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,128,0.014122666584120857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,128,0.00508711114525795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,3072,0.007601777712504069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,5120,0.01811200049188402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,12288,0.0351182222366333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,768,0.015838222371207345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,128,0.01482399966981676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,2560,0.026475555366939966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,8192,0.15265156163109675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,16384,0.062048892180124916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,128,0.01407999959256914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,2048,0.19371467166476777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,3584,0.011984888878133563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,12288,0.02440266642305586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,1536,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,16384,0.04937422275543213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,128,0.01756800048881107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,8192,0.04179555508825514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,7168,0.011389333340856763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,512,0.013776889277829064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,16384,0.09004088905122544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,3584,0.013140444126394061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,2048,0.006622222148709827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,256,0.01685688893000285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,256,0.006994666324721442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,7168,0.045075555642445884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,2048,0.011772444678677453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,4096,0.007583111524581909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,512,0.003168888803985384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,10240,0.012392889294359418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,7168,0.08127822478612264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,768,0.004202666795916027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,2560,0.015467555986510383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,16384,0.03464444478352865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,10240,0.01681688924630483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,5120,0.013452444639470843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,10240,0.01275199982855055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,8192,0.1525742212931315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,32,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,10240,0.02000266644689772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,768,0.01408622165520986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,16384,0.04351733459366692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,32,0.0020977778153287037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,768,0.016137777103318107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,65536,0.6081902186075846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,128,0.002499555548032125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,7168,0.03399111164940728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,5120,0.013752000199423896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,4096,0.0099982221921285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,1024,0.003181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,4096,0.016089777151743572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,2560,0.006791111081838608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,3584,0.01408622165520986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,64,0.007600888609886169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,256,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,4096,0.06455022096633911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,10240,0.7715786827935113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,2560,0.014583110809326172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,256,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,8192,0.047064890464146934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,6144,0.009006222089131674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,2560,0.015478221906556023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,128,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,3072,0.018744889232847426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,64,0.0021359999146726397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,7168,0.01495911180973053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,16384,0.015832889411184523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,128,0.014547554983033074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,3072,0.025116443634033203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,65536,0.0836755567126804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,8192,0.024526221884621516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,32,0.001964444501532449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,768,0.011112000379297467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,1024,0.014877332581414117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,5120,0.07671111159854464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,512,0.00535111129283905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,8192,0.07540800174077351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,65536,0.12038311693403457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,12288,0.005621333503060871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,1024,0.006296000132958095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,8192,0.011071110765139261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,65536,0.07522222068574694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,64,0.002230222233467632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,65536,0.08448888858159383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,5120,0.008651555412345463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,16384,0.22339023484124076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,2560,0.12285955746968587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,2560,0.004907555464241239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,8192,0.01072355525361167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,6144,0.1255964438120524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,128,0.00217333332531982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,1024,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,4096,0.01278133359220293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,6144,0.018260444204012554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,6144,0.019912888606389362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,10240,0.035622222556008235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,128,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,5120,0.040157334672080144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,8192,0.05384977658589681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,7168,0.05091555582152473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,65536,0.16484622160593668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,512,0.002532444480392668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,2048,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,65536,0.07748177978727552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,2560,0.004534222185611725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,12288,0.1886204481124878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,2048,0.02865866820017497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,256,0.022408889399634466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,3072,0.06740799877378675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,2048,0.006573333508438534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,256,0.002185777864522404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,7168,0.07697244485219319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,512,0.005583111196756363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,7168,0.051422221793068774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,10240,0.01512888901763492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,10240,0.017160000072585214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,10240,0.025260445144441392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,7168,0.019932443896929424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,2560,0.008295111358165741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,32,0.001818666648533609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,7168,0.01645155582163069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,3584,0.015791111522250705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,1024,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,3584,0.016903999778959487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,10240,0.06370311313205294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,65536,0.11395111348893906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,256,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,3584,0.020967110991477966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,12288,0.007294221884674496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,4096,0.04945244391759237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,512,0.014834667245546976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,12288,0.18993688954247367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,768,0.00591111100382275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,1536,0.015572445260153877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,3072,0.007175111108356052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,4096,0.36577155854966903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,8192,0.018703111343913607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,256,0.015855999456511605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,10240,0.0812986691792806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,6144,0.04282044370969137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,6144,0.02686133318477207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,16384,0.05206400156021118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,65536,0.1238586637708876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,2560,0.008661333057615492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,256,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,256,0.0069431112044387394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,4096,0.09858666525946723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,4096,0.008411555654472774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,65536,0.2744853231641981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,12288,0.022648889157507155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,64,0.007328888608349695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,1536,0.00795288880666097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,256,0.01237777786122428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,64,0.005437333136796951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,2048,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,2048,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,768,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,1024,0.014490667316648694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,768,0.00832266691658232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,256,0.012727999852763282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,512,0.0032017777363459268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,512,0.0053182223604785065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,1536,0.033568888902664185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,8192,0.054245332876841225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,3072,0.007639110916190677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,12288,0.28883110152350533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,16384,0.07150666581259833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,32,0.0052888890107472735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,32,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,768,0.003867555409669876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,3584,0.017514665921529133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,65536,0.07510311073727079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,4096,0.00866844422287411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,8192,0.01071466671095954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,2560,0.007621333003044128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,8192,0.011026666396194033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,3072,0.011773332953453064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,5120,0.008278222547637092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,2048,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,3584,0.011392888923486074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,12288,0.059614221254984535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,3584,0.008808000220192803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,8192,0.0696835584110684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,2560,0.03449511196878221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,6144,0.040008889304267034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,16384,0.7778133286370171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,32,0.0022213332768943575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,3072,0.009004444711738164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,2048,0.015388444066047668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,1536,0.006721777634488211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,128,0.005071111023426056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,1536,0.003607999947335985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,2048,0.008812444077597724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,8192,0.012496888637542725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,7168,0.04565511147181193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,10240,0.011328889264000786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,64,0.00795022232664956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,3584,0.04104177819357978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,32,0.04607466525501675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,65536,0.06536799669265747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,10240,0.00996888925631841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,6144,0.023883556326230366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,1024,0.01568533314598931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,768,0.006981333096822103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,2560,0.027564444475703772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,16384,0.02476177778508928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,256,0.002858666703104973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,3072,0.007292444507280986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,512,0.010130667024188572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,3584,0.02602311140961117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,16384,0.02650755478276147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,5120,0.03664355476697286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,768,0.0058613332609335584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,7168,0.015472888946533203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,128,0.01311644415060679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,256,0.0031582222630580268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,1536,0.027075555589463975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,10240,0.03628355595800612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,1024,0.0031786666562159858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,1536,0.005939555664857228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,256,0.016720000240537856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,1024,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,16384,0.03251288996802436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,3584,0.013563555147912769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,3072,0.07345777750015259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,65536,0.08559822373920017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,768,0.0052586665583981415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,768,0.008623999853928884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,8192,0.1335102187262641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,768,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,2048,0.006559111177921295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,128,0.013766222529941134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,32,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,1024,0.01312444441848331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,1024,0.00574311117331187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,4096,0.00832266691658232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,3584,0.010983111129866706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,16384,0.03398844599723816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,32,0.014440000057220459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,32,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,512,0.02126844392882453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,7168,0.013694222602579327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,12288,0.024294222394625347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,7168,0.018935110833909776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,10240,0.01462222304609087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,256,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,6144,0.031790223386552595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,3072,0.007678222325113084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,6144,0.025906667113304138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,32,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,1024,0.003119999956753519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,512,0.004898666507667965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,5120,0.008967111508051554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,64,0.002483555633160803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,1536,0.08721511231528388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,1024,0.010759111079904767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,4096,0.020120888948440552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,3072,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,4096,0.015674667225943673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,6144,0.020759110649426777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,7168,0.6805529064602323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,2048,0.0439644455909729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,256,0.003181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,256,0.02641600039270189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,2560,0.02053955528471205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,7168,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,128,0.011029333704047732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,8192,0.04084622197681003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,2560,0.006647999915811751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,7168,0.0331582228342692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,7168,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,10240,0.23237599266899955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,1536,0.0047208890318870544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,768,0.010388444695207808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,3584,0.004575110971927643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,4096,0.006650666809744305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,5120,0.018191110756662156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,65536,0.07853866948021783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,64,0.001975999938117133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,128,0.013909333282046847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,7168,0.010120888551076254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,1536,0.006099555641412735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,5120,0.11124444007873535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,2560,0.007214222517278459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,3072,0.00923555592695872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,4096,0.021242666575643752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,10240,0.02718222141265869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,12288,0.01110222190618515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,3584,0.008580444587601556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,10240,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,10240,0.016172443826993305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,2560,0.006038222048017714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,2048,0.030517333083682593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,768,0.005879111174080107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,2560,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,6144,0.016873778568373788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,65536,0.17286133766174316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,1024,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,1536,0.015014222926563688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,5120,0.009282666775915358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,64,0.023151111271646287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,3072,0.023746665981080797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,3584,0.015441776977645027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,768,0.014483556151390076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,1024,0.050247109598583646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,16384,0.026956443985303242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,7168,0.16462488969167074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,16384,0.1430471075905694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,3072,0.007983999947706858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,2048,0.7403013441297742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,12288,0.1861813333299425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,1024,0.00425511101881663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,16384,0.07373422384262085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,5120,0.019512888458040025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,10240,0.01904533306757609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,768,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,1024,0.01687111126052009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,3584,0.010284444524182213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,8192,0.017863111363516915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,1536,0.0058746664888328975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,8192,0.029233776860766943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,64,0.0021359999146726397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,5120,0.0370488895310296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,2560,0.031108445591396753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,3072,0.021556443638271753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,64,0.006671111202902264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,8192,0.028831998507181805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,3584,0.010398222340477837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,8192,0.020703110429975722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,7168,0.0053288886944452924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,768,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,12288,0.019160888261265226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,6144,0.1293964385986328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,2048,0.014561777313550314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,8192,0.08470311429765488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,32,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,4096,0.008667555948098501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,32,0.0018453333112928604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,2048,0.01789422167672051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,3072,0.1414817836549547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,32,0.005610666755172942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,2560,0.015141333142916361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,8192,0.010320888625250923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,512,0.00656088896923595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,7168,0.03681688838534885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,1536,0.013080889152155982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,512,0.003846222327815162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,128,0.0028008888992998335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,4096,0.005192888693677055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,8192,0.011699555648697747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,8192,0.021340444684028625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,256,0.013786666923099093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,3584,0.009335111412737105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,128,0.01348355578051673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,1024,0.02513955533504486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,65536,0.21900622049967447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,128,0.017107556263605755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,256,0.009352888498041365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,128,0.007798222204049428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,12288,0.02295555505487654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,128,0.006618666566080517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,2560,0.01645955608950721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,3072,0.00833777752187517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,2048,0.010603555374675326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,256,0.016194666425387066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,256,0.006647111227114995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,5120,0.03839377893341912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,768,0.010915555887752108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,16384,0.0281644430425432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,8192,0.016283555163277518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,7168,0.030232889784706965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,3072,0.007599111232492659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,512,0.0028853333658642242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,32,0.0022604444788561925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,64,0.005963555640644497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,65536,0.05851022402445475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,12288,1.2845537397596571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,512,0.015501333607567681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,128,0.005680888891220093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,4096,0.020198222663667467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,2560,0.008965333302815756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,1024,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,3584,0.01107911103301578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,2560,0.007015110717879401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,16384,0.037216888533698186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,6144,0.029834667841593426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,768,0.003528888854715559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,128,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,65536,0.04588355620702108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,6144,0.44306490156385636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,2560,0.029134223858515423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,512,0.005863111052248213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,2560,0.026072000463803608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,64,0.007290666302045186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,12288,0.03914755582809448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,512,0.0029031110720502008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,2560,0.007227555745177799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,7168,0.010043555663691627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,3584,0.030910223722457886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,6144,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,2560,0.01921955578856998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,512,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,7168,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,128,0.005528000079923206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,512,0.0028853333658642242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,512,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,2560,0.013445333474212222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,3584,0.015903110305468243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,128,0.046386665768093534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,256,0.22736710972256133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,1024,0.015837333268589444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,1024,0.014361777239375643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,12288,0.1414168940650092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,256,0.015588444140222339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,6144,0.011744889120260874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,128,0.032840887705485024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,6144,0.03195377853181627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,12288,0.04158044523662991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,768,0.006638222270541721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,8192,0.025713778204388086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,65536,0.058463109864128955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,128,0.011030221978823343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,2560,0.02530133393075731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,10240,0.020032889313167997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,1024,0.014567999376191033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,5120,0.08786222007539536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,32,0.02331200076474084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,10240,0.047510223256217114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,2048,0.5959217813279893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,5120,0.014364444547229342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,3072,0.010337777435779572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,3072,0.01680266691578759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,1536,0.006583999842405319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,64,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,65536,0.12676089339786106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,8192,0.04387199878692627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,4096,0.00865955568022198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,2048,0.00665244460105896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,1024,0.007328888608349695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,512,0.005666666560702854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,3584,0.037280887365341187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,1536,0.009993777506881291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,5120,0.008623999853928884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,128,0.0021804444905784395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,64,0.0018444444156355327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,2560,0.008956444760163626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,3072,0.05041244294908312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,16384,0.02482044365670946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,16384,0.017860444055663217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,7168,0.025731555289692346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,8192,0.028775999943415325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,128,0.014167999227841696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,1024,0.0028479999552170434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,256,0.008357333640257517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,65536,0.5126177999708388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,12288,0.024676443801985845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,128,0.014451555079884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,8192,0.01037599974208408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,4096,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,2560,0.009659555223253038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,1536,0.0038542221817705366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,512,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,8192,0.023668444818920557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,512,0.006627555522653792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,256,0.009366222553782994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,10240,0.06790577703052096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,4096,0.01313511116637124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,10240,0.08235288990868463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,32,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,1024,0.06986577643288507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,8192,0.018585777944988675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,1536,0.018184888694021437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,8192,0.00553688903649648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,2560,0.006919999917348226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,2560,0.016220443778567843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,4096,0.04348444607522753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,16384,0.00573600000805325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,7168,0.07492444250318739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,4096,0.025753777888086107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,1536,0.01056177748574151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,5120,0.04287022352218628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,3072,0.01424711114830441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,768,0.0028195555011431375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,768,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,10240,0.03461777832772996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,2560,0.014141332772043018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,4096,0.03203644355138143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,256,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,8192,0.02342222299840715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,256,0.048400001393424145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,3584,0.007289778027269576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,8192,0.01255733354224099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,2560,0.0120328888297081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,16384,0.02233422299226125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,64,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,8192,0.0052586665583981415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,32,0.006232000059551663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,1024,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,1536,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,8192,0.013019555144839816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,2048,0.011021333436171213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,768,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,65536,0.024736889534526404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,128,0.014091556270917257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,1024,0.00425511101881663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,4096,0.008400888906584846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,16384,0.04769066638416714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,128,0.014858666393491956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,65536,0.14057066705491808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,256,0.02681066592534383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,128,0.03001333276430766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,32,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,65536,0.10289244519339667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,65536,0.07080977492862277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,3584,0.015522667103343539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,256,0.002473777780930201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,3072,0.06778399811850654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,4096,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,65536,2.675407197740343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,10240,0.02474133339193132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,128,0.009626666704813639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,12288,0.020541333489947848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,128,0.015481778317027621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,10240,0.011412444214026133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,4096,0.02308533257908291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,3072,0.008738666772842407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,10240,0.02641244398223029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,6144,0.011253332926167382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,2560,0.00869511150651508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,32,0.002435555474625693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,2560,0.1611057784822252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,3584,0.016373333003785875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,4096,0.00452622233165635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,512,0.2747253312004937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,3584,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,7168,0.010341333018408881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,3072,0.006629333313968446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,10240,0.05845866600672404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,1536,0.010434666441546546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,768,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,3072,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,6144,0.009395555489593083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,768,0.0037164443896876443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,256,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,10240,0.015672889020707872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,7168,0.014194667339324951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,2560,0.015776889191733468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,128,0.004935111022657818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,512,0.01684177749686771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,10240,0.016294222739007738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,512,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,3584,0.015114666687117683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,12288,0.008307555483447181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,3072,0.01384622189733717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,128,0.0021946666141351065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,3072,0.02932888931698269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,5120,0.058965331978268094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,1024,0.004888888862397936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,3584,0.04472000069088406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,128,0.005251555393139522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,32,0.007319110963079665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,4096,0.007974222302436829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,8192,0.037816001309288874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,3072,0.014224888549910652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,3584,0.011774222056070963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,1024,0.008349333372380998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,3072,0.015424888994958667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,256,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,6144,0.027032888597912256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,6144,0.009353777600659264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,5120,0.016943110360039603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,2560,0.017342221405771043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,2560,0.010513777534166971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,7168,0.04690133200751411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,1536,0.0069057775868309875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,768,0.00488711107108328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,10240,0.011808888779746162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,3584,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,7168,0.053654223680496216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,768,0.00544799988468488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,3072,0.01034222212102678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,5120,0.009290667043791877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,65536,0.06235644552442762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,768,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,768,0.0036071112586392295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,32,0.005643555687533484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,1536,0.006326222171386083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,512,0.0029128889242808023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,7168,0.026055110825432673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,10240,0.05088355475001865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,65536,0.06777777936723497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,65536,0.05446222093370226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,128,0.004938666605287128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,2560,0.0075031113293435835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,512,0.014153778553009033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,3584,0.013434666726324292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,256,0.014144889182514615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,2048,0.014825777875052558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,128,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,1536,0.024884444144037034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,65536,0.15333955817752415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,3072,0.012044444680213928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,6144,0.0099653336736891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,2560,0.00833422193924586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,2560,0.01788444485929277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,128,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,10240,0.03852533300717672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,1024,0.002887111157178879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,3584,0.018198221921920776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,2560,0.005395555661784277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,2048,0.006592888798978593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,3072,0.018033777674039204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,2560,0.010052444206343757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,64,0.005888888819350137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,12288,0.034303112162484065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,10240,0.011744000017642975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,1024,0.005965333431959152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,1024,0.43307110998365617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,2048,0.015180443723996481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,32,0.002575111058023241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,1536,0.01051288925939136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,16384,0.1154435608122084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,1024,0.016162667009565566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,4096,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,1536,0.013000888956917657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,12288,0.024567999773555334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,1536,0.006596444381607904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,5120,0.0199306673473782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,3584,0.010753778119881948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,32,0.007616000043021307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,12288,0.01404355631934272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,7168,0.012343111137549082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,5120,0.012777778009573618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,512,0.020495999190542433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,256,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,10240,0.04732799861166212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,768,0.011715555356608497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,5120,0.013919111755159167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,3072,0.007778666913509369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,1536,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,1024,0.00453244439429707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,5120,0.020863110820452373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,65536,0.05843466520309448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,128,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,1536,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,12288,0.013815999858909182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,4096,1.2354915406968858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,256,0.006604444649484422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,6144,0.020576889316240948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,6144,0.028411553965674505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,1536,0.006220444622966979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,16384,0.04025955663786994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,7168,0.010000000397364298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,768,0.01519377695189582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,5120,0.12354132864210342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,2048,0.006629333313968446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,128,0.005047111047638787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,4096,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,1024,0.005888888819350137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,12288,0.09852533207999335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,2560,0.006640000061856375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,12288,0.018968888454967074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,7168,0.010025777750545079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,5120,0.015032889114485847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,6144,0.02404888967672984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,128,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,10240,0.023022222850057814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,2048,0.01889955500761668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,5120,0.01275466630856196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,3584,0.007983999947706858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,3584,0.010411555568377176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,65536,0.04470666580730014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,7168,0.005292444593376583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,256,0.0022453332526816265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,1536,0.03934577769703335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,3072,0.06912355290518866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,256,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,32,0.0038720000949170855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,3072,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,768,0.02605777813328637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,4096,0.021849777963426378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,32,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,16384,0.019339554839664035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,12288,0.020637333393096924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,512,0.0240346673462126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,32,0.0019128889673286015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,32,0.006055999961164262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,2560,0.007419555551475949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,6144,0.011176889141400656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,10240,0.054458667834599815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,512,0.027670222851965163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,768,0.018704889549149405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,5120,0.008623999853928884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,5120,0.06300444735421075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,32,0.011236444115638733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,512,0.02197422252760993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,3584,0.004278222305907143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,5120,0.16509955459170872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,3072,0.9368933571709527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,8192,0.015542222393883599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,128,0.29598932796054417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,12288,0.026760000321600173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,10240,0.014113777213626437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,768,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,1536,0.012715555727481842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,2560,0.006911110960774952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,128,0.004928888960017098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,32,0.016155555844306946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,2560,0.007648000286685095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,128,0.005365333209435145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,2048,0.00628977765639623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,16384,0.07802489068773058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,1536,0.01860000027550591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,16384,0.12934133741590711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,65536,0.21786755985683867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,6144,0.037039111057917275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,2560,0.01720888912677765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,16384,0.03016977839999729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,7168,0.013441777891582914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,7168,0.013727111121018728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,8192,0.01994400057527754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,16384,0.02092622220516205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,512,0.002587555597225825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,4096,0.03658311234580146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,3584,0.036121779017978244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,1024,0.09770133097966512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,10240,0.049676444795396596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,6144,0.00977244476477305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,10240,0.3311857647365994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,768,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,1024,0.014595554934607612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,512,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,768,0.006574222197135289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,4096,0.0203164451652103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,7168,0.01790844400723775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,65536,0.026485333840052288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,768,0.002842666581273079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,3584,0.014874666929244995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,6144,0.009296000003814697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,1536,0.006579555571079254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,12288,0.04369689027468363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,6144,0.50017335679796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,6144,0.03442310955789354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,3072,0.006573333508438534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,2048,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,3072,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,12288,0.017215111189418368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,768,0.00424799985355801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,1024,0.023224888576401606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,1024,0.0036880001425743103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,2048,0.00612088872326745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,8192,0.02234399980968899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,7168,0.04669777883423699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,7168,0.1489528947406345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,12288,0.037857777542538114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,4096,0.013108444710572561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,7168,0.10558399889204238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,5120,0.12496444914076064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,2048,0.01183377785815133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,16384,0.006630222416586346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,3072,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,3584,0.26758310529920787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,6144,0.025774222281244066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,256,0.002507555608948072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,4096,0.008394666843944127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,7168,0.009698666632175446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,7168,0.01977688901954227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,3072,0.032467554012934365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,1536,0.006623999940024481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,512,0.005892444401979446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,5120,0.012524444195959302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,3584,0.0074622225430276655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,1536,0.006775110960006714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,2048,0.018173333671357896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,8192,0.01795644395881229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,256,0.014500444134076437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,512,0.0052888890107472735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,6144,0.018193778064515855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,3072,0.01754755609565311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,4096,0.014847111370828418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,16384,0.07935911417007446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,4096,0.09191289212968613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,768,0.018076444665590923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,1536,0.06193421946631538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,32,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,5120,0.01145511120557785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,3584,0.011175110936164856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,7168,0.058281779289245605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,128,0.006329777754015393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,7168,0.023967999551031325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,512,0.01830844415558709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,3584,0.10101777977413601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,64,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,12288,0.014045332868893942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,128,0.06003555324342516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,4096,0.008549333446555668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,8192,0.01107022249036365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,65536,0.08233510785632663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,512,0.010760889285140567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,2048,0.009328888522254096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,3072,0.01500266624821557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,128,0.0030213333666324615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,16384,0.06345511145061918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,7168,0.18729510572221544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,12288,0.02735022207101186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,3072,0.014448889427714877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,6144,0.029112001260121662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,4096,0.007914666500356462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,64,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,256,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,65536,0.16265866491529676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,3072,0.030814223819308813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,65536,0.089610669347975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,12288,0.0141004439857271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,16384,0.07853155665927462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,10240,0.012900444368521372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,16384,0.5291057692633735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,4096,0.013814222481515674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,6144,0.009331555830107795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,65536,0.06323111057281494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,3072,0.017799999978807237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,3584,0.009201777478059133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,2048,0.006636444479227066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,32,0.017571555243598092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,4096,0.011095999843544431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,3072,0.0069164443347189165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,16384,0.025755556093321905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,7168,0.021228444245126512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,768,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,65536,0.31538934177822536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,4096,0.011060444845093621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,1536,0.009000889129108852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,6144,0.012048888537618848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,128,0.012708444562223224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,256,0.005203555441564984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,12288,0.1434817843967014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,32,0.009663110805882348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,65536,0.06381777922312419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,1536,0.00794933322403166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,8192,0.021200888686709937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,65536,0.26765065722995335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,1536,0.007600888609886169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,65536,0.05948533614476522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,7168,0.02200266718864441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,3584,0.09551199939515854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,65536,0.9144435458713107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,5120,0.01404088901148902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,10240,0.014096889230940076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,256,0.009571555587980483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,2048,0.029459555943806965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,256,0.0236151119073232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,2048,0.004256000121434529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,12288,0.4034630987379286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,65536,0.14049332671695286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,10240,0.25498223304748535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,7168,0.05116000109248691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,4096,0.1306026644176907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,10240,0.22683111826578775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,3584,0.016003555721706815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,5120,0.021151110529899597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,16384,0.28111910820007324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,16384,0.056806219948662647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,3072,0.006649777707126405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,12288,0.022308443983395893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,6144,0.012025777664449481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,256,0.016206221448050607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,10240,0.26497689882914227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,1536,0.005967999911970562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,2048,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,512,0.013807111316257052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,12288,0.03667022122277154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,768,0.00592533333433999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,3072,0.016976000534163583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,65536,0.020824000239372253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,5120,0.031256887647840716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,12288,0.019696000549528334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,512,0.026456000076399908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,7168,0.03987733191914029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,768,0.005942222144868638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,5120,0.015463110473420886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,3584,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,768,0.003133333391613431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,3584,0.007320000065697565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,3072,0.015820443630218506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,2048,0.12695466147528753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,2048,0.09198133150736491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,512,0.02131822208563487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,5120,0.008295111358165741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,256,0.015556444724400839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,1024,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,1536,0.007968000239796108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,5120,0.021360889077186584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,5120,0.009331555830107795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,6144,0.01635199950801002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,7168,0.004984000076850255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,512,0.006586666736337874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,7168,0.06403555472691853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,3584,0.011477332976129321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,10240,0.011776000261306763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,12288,0.059225777784983315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,1536,0.005982222242487802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,5120,0.017729777428838942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,4096,0.038359999656677246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,16384,0.015099555253982544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,7168,0.009701333112186855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,3072,0.013599110974205865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,3584,0.007600888609886169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,3072,0.007315555380450354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,1536,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,5120,0.01847466660870446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,4096,0.00998133338159985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,3584,0.08853155374526978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,2560,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,1024,0.017543110582563613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,2560,0.005968889014588461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,12288,0.02039377722475264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,128,0.005305777821275923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,256,0.003903999924659729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,2560,0.012702222499582501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,10240,0.012434666355450949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,12288,0.01276177747382058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,4096,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,512,0.0029066666546795103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,3584,0.014416888356208801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,2560,0.024322667055659827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,256,0.009346666435400644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,8192,0.0120666664507654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,2048,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,16384,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,7168,0.014869333969222175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,512,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,3584,0.007691555553012424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,16384,0.010678222609890832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,2560,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,128,0.004549333204825719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,512,0.014454222387737699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,5120,0.012039999994966718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,1024,0.03609155615170797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,4096,0.013578666581047906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,8192,0.012085333466529846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,10240,0.26251912117004395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,3584,1.210004488627116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,16384,0.025293333662880793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,3072,0.02629599968592326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,3072,0.009322666459613377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,2560,0.004917333523432414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,512,0.004936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,6144,0.009011555876996782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,1536,0.0035013332962989807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,6144,0.009319110876984065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,10240,0.006636444479227066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,256,0.01477599971824222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,2048,0.004801777915822135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,512,0.005321777943107817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,12288,0.017140444782045152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,256,0.00452622233165635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,6144,0.009691555466916826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,768,0.07081244389216106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,12288,0.02198133369286855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,65536,0.20456355147891572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,3072,0.007630222373538547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,3072,0.02295822236273024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,65536,0.10073155826992458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,12288,0.014095111025704278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,512,0.006976000136799282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,6144,0.012075555821259817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,6144,0.04500977860556709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,16384,0.041943109697765775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,256,0.021380444367726643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,5120,0.02440533373090956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,10240,0.015838222371207345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,7168,0.016528000434239704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,65536,0.3280959924062093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,10240,0.005063999858167437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,2560,0.10068533155653213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,5120,0.008472888833946651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,1024,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,2560,0.00637155564294921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,7168,0.02789155642191569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,7168,0.2616906695895725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,1536,0.006344888773229387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,10240,0.020682666036817763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,6144,0.344127098719279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,64,0.002157333410448498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,8192,0.015772443678643968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,2560,0.004865777575307422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,768,0.015490666031837463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,10240,0.06352266338136461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,2048,0.026483555634816486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,12288,0.021357332666714985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,10240,0.012034667034943899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,2560,0.0240320000383589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,768,0.005616888817813661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,64,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,2560,0.00868088917599784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,1536,0.03277955452601115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,16384,0.041069332096311785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,1024,0.013090666797426013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,256,0.0028391112056043413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,4096,0.014164444473054675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,128,0.002135111019015312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,2048,0.006270222365856171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,12288,0.019355555375417072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,6144,0.04535377687878079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,5120,0.016827555166350473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,128,0.011011555790901184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,16384,0.016573333077960543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,256,0.01664355562792884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,3072,0.02567822237809499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,2560,0.005193777796294955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,1024,0.013207111093733044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,3584,0.007452444069915348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,3584,0.00869066682126787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,32,0.0019013333237833446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,12288,0.021809778279728357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,256,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,512,0.011888888974984487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,256,0.013364444176355997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,3584,0.020411555965741474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,256,0.04354755414856804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,3584,0.007651555869314406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,4096,0.013034666577974955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,512,0.005677333308590784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,8192,0.053363555007510714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,64,0.005648889061477449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,3584,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,12288,0.014112000664075216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,8192,0.0700391133626302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,2048,0.046318223079045616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,6144,0.01312444441848331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,2560,0.02493511140346527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,3072,0.006962666908899943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,16384,0.032933334509531655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,4096,0.008918222453859117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,2048,0.11475200123257107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,5120,0.055016001065572105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,10240,0.014941333068741692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,768,0.003910222401221593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,16384,0.019545776976479422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,10240,0.03674666749106513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,10240,0.03155644403563605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,2048,0.008273777862389883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,16384,0.026104888982243005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,7168,0.017814222309324477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,32,0.005706666658322017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,10240,0.017255999975734286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,512,0.012131555212868584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,256,0.005816000203291575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,512,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,1536,0.02770044406255086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,8192,0.01072177787621816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,4096,0.026897778113683064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,1536,0.018921777606010437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,65536,0.06739022334416707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,256,0.002830222249031067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,4096,0.009914666414260864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,8192,0.017695999807781644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,12288,0.013806222213639153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,16384,0.032920890384250216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,2560,0.028498666154013738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,128,0.014117333624098035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,5120,0.020446222689416673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,8192,0.02187999917401208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,2048,0.008905777500735389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,7168,0.009663999908500248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,4096,0.04745422138108147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,4096,0.03827555643187629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,7168,0.05035822259055244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,6144,0.010104888843165504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,12288,0.12107200092739528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,128,0.01517777807182736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,512,0.013432888521088494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,5120,0.031184001101387873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,768,0.017451556192504037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,2560,0.03980266716745164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,12288,0.02870933214823405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,8192,0.03978311022122701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,4096,0.017988445030318368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,7168,0.01806400054030948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,1536,0.005953777581453323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,2048,0.032046222024493747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,12288,0.12896445062425402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,1024,0.011074666347768573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,10240,0.01718222267097897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,65536,0.10926310883627997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,12288,0.028566221396128338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,128,0.0021377777059872947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,3072,0.009810667071077559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,768,0.00831822223133511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,2560,0.018751111295488145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,2048,0.01997955474588606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,12288,0.15505954954359266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,1024,0.011411555939250521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,1024,0.003141333245568805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,5120,0.008379555410808986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,7168,0.10908533467186822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,2048,0.0185253338681327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,7168,0.013213333156373767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,2560,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,32,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,512,0.04084800018204583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,3072,0.014780445231331719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,8192,0.0155395550860299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,16384,0.01683644453684489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,7168,0.02091288897726271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,768,0.021888888544506494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,7168,0.5312986903720432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,3584,0.007314666277832455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,10240,0.16345955265892878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,768,0.0069066666894488865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,5120,0.004181333300140169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,65536,0.07553778092066447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,256,0.014883556299739413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,12288,0.029755555921130713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,1024,0.005671999934646819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,512,0.016819554898473952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,1536,0.01295199990272522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,4096,0.00832177781396442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,8192,0.012777778009573618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,512,0.016903999778959487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,512,0.0025635556214385563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,65536,0.12307733959621853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,1024,0.006296000132958095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,32,0.0039057777159743835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,5120,0.019107555349667866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,7168,0.005306666509972678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,512,0.005877333382765452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,10240,0.03106844425201416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,2560,0.08032888836330838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,4096,0.00851111114025116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,2048,0.006579555571079254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,768,0.0033191111352708605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,6144,0.01885422236389584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,512,0.00796977761718962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,1536,0.005961777849329843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,3584,0.032744000355402626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,7168,0.014845333165592618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,8192,0.010659555594126383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,32,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,768,0.00589155571328269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,3584,0.00794311116139094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,256,0.015080889066060385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,8192,0.01391377713945177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,6144,0.011468444433477191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,2560,0.01586933268441094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,4096,0.011917333636018964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,7168,0.018585777944988675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,6144,0.014353776971499125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,256,0.003528888854715559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,4096,0.016157334049542744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,768,0.0332551101843516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,768,0.006128888991143968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,12288,0.020629333125220407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,8192,0.046160889996422656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,6144,0.014270222849316068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,5120,0.009675555759006077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,10240,0.07101688782374065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,32,0.001791999985774358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,2560,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,1536,0.00907733374171787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,3584,0.008022222254011366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,65536,0.13939289251963297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,256,0.0028826666788922418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,6144,0.6818115446302625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,5120,0.011376000112957425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,1536,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,1536,0.01575377749072181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,32,0.002193777718477779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,128,0.008626666333940294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,8192,0.020594666401545208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,12288,0.019528888993793063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,4096,0.005945777727497949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,8192,0.0590826670328776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,1024,0.0031679999083280563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,65536,0.07987822426689996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,6144,0.10283466842439438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,10240,0.020579554968410067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,7168,0.3071982330746121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,2560,0.019644444187482197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,768,0.005528000079923206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,12288,0.08265688684251574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,128,0.014000000225173103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,4096,0.012686221963829465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,2048,0.03660533494419522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,256,0.004884444591071871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,128,0.020224000016848247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,1536,0.005985777825117111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,16384,0.1723804473876953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,1024,0.01647555496957567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,128,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,512,0.005933333188295364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,12288,0.014759999182489185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,1536,0.0038942222793896994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,1024,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,768,0.03546844588385688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,512,0.016181333197487723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,128,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,64,0.007793777518802219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,6144,0.009955555200576782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,12288,0.01830666595035129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,6144,0.025968889395395916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,2560,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,4096,0.011733333269755045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,3072,0.013979555832015144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,4096,0.007643555601437886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,3072,0.01038577738735411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,65536,0.08701599968804254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,3072,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,1024,0.021319111188252766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,4096,0.003859555555714501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,10240,0.07458400064044528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,2048,0.02776266634464264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,2048,0.004586666822433472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,768,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,3584,0.007988444632954067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,4096,0.008992888861232335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,65536,0.02738933265209198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,64,0.022319111559126113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,10240,0.01993688941001892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,10240,0.07015555434756808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,2560,0.006849777781301075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,512,0.013787555197874704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,256,0.012806221842765808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,3072,0.013431111143694984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,2048,0.006606222026877933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,6144,0.341447114944458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,2560,0.006912888752089606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,10240,0.02036622166633606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,2048,0.0042168887125121225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,3584,0.03728533453411526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,3584,0.024033778243594702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,1536,0.005128889034191768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,12288,0.021773333350817364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,7168,0.005205333232879639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,4096,0.02012711101108127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,16384,0.026280888252788123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,512,0.0028275555620590844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,5120,0.04549333453178406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,4096,0.011875555747085147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,6144,0.010645333263609143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,512,0.005872888697518243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,16384,0.1726764440536499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,4096,0.11643732918633355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,3072,0.11424622270796035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,256,0.006568888823191325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,12288,0.024226667152510747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,10240,0.022544888986481562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,512,0.005436444448100195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,6144,0.010724444356229572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,3072,0.021533333592944678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,7168,0.03494755427042643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,1536,0.006763555523422029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,3584,0.02277066641383701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,3072,0.022800889280107286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,3072,0.006183111005359226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,10240,0.024501333634058636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,6144,0.006218666831652324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,512,0.012063999970753988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,12288,0.07082400057050917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,3072,0.01311288856797748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,3584,0.01814133259985182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,32,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,2048,0.004594666676388847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,32,0.0058702222175068325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,768,0.011752000285519494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,8192,0.054152889384163745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,1024,0.013227555486891003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,12288,0.04050666756100125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,12288,0.0052871112194326185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,3072,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,2560,0.017512889371977914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,512,0.011339555184046427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,12288,0.016679111454221938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,7168,0.041872888803482056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,32,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,12288,0.06576710939407349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,10240,0.022081777453422546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,64,0.005976000179847081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,3584,0.029789331886503432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,256,0.005205333232879639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,1024,0.04910222358173794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,4096,0.027886221806208294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,16384,0.044320887989468045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,12288,0.03542666633923849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,8192,0.038520889149771795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,65536,0.05563555823432075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,10240,0.036687999963760376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,2560,0.10382399956385295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,512,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,128,0.0035582222044467926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,8192,0.017487999465730455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,3072,0.01072266697883606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,1536,0.005954666684071223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,64,0.001863111017478837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,1536,0.0047324444684717385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,2560,0.009323555562231276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,1024,0.016524444023768108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,768,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,1024,0.003018666679660479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,3072,0.006978666616810693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,1024,0.050944000482559204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,10240,0.012429333395428128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,512,0.012457778056462606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,64,0.011396444506115384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,7168,0.07488266626993816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,1024,0.08398755391438802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,1536,0.026294221480687458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,3072,0.007263110743628607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,2048,0.017436444759368896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,16384,0.0291608903143141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,1536,0.02360711163944668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,3584,0.008339555727110969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,512,0.003896000070704354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,2560,0.02421244482199351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,8192,0.05015200045373705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,7168,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,10240,0.04853599932458666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,768,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,6144,0.009682666924264696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,5120,0.012118221984969245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,6144,0.01443199978934394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,128,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,3584,0.030064887470669214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,10240,0.024215110474162634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,1024,0.01585422290696038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,1536,0.022999111149046157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,5120,0.011408888631396823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,4096,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,10240,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,4096,0.005646222167544895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,12288,0.022971555590629578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,1536,0.02568799919552273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,256,0.0021271111650599376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,1024,0.003048000029391713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,12288,0.022292445103327434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,512,0.05851022402445475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,10240,0.01700888905260298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,512,0.011775111158688864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,32,0.0028737777223189673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,512,0.013068444199032254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,3584,0.020964443683624268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,4096,0.008963555925422246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,3584,0.01276533305644989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,4096,0.03401244348949856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,1536,0.01037599974208408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,2560,0.22907466358608672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,12288,0.02990844514634874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,2048,0.0046862223082118565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,10240,0.01240799989965227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,65536,0.0870728890101115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,16384,0.08052888843748304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,512,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,1536,0.04117688867780898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,12288,0.01442488862408532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,5120,0.08243022362391154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,4096,0.06025333536995781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,8192,0.027454222242037456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,4096,0.008991111483838823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,3072,0.03225422236654494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,8192,0.011021333436171213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,8192,0.13830133279164633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,10240,0.06304533614052667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,256,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,512,0.021632888250880774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,64,0.013398222625255585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,10240,0.016843555702103507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,64,0.0053395554423332214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,256,0.005175999883148405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,6144,0.02156177825397915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,65536,0.08434400293562148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,12288,0.013824000126785703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,2560,0.038570665650897555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,3584,0.013656888571050433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,3072,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,6144,0.007711110843552484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,7168,0.06695822212431166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,6144,0.037998222642474704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,65536,0.1212728950712416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,8192,0.01073333372672399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,32,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,32,0.002185777864522404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,6144,0.013807111316257052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,256,0.003237333355678452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,12288,0.021175111333529156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,8192,0.0103511114915212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,256,0.009020444419648912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,2560,0.014025777578353882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,1024,0.004265777766704559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,4096,0.00849599970711602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,768,0.015623999966515435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,2560,0.006913777854707506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,3584,0.010752889017264048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,768,0.007304888632562425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,2560,0.006640888750553131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,64,0.007332444190979004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,65536,0.027838223510318335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,256,0.0022213332768943575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,12288,1.0030071470472548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,128,0.015850666496488783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,3584,0.01034400032626258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,64,0.0018399999373488957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,16384,0.034006221426857844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,12288,0.0281368891398112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,5120,0.014465777410401238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,65536,0.18169332875145805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,5120,0.019056000643306308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,768,0.0032195556494924757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,2048,0.009383999639087254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,5120,0.013369777964221107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,128,0.003871110992299186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,32,0.004861333303981357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,5120,0.024642666180928547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,2048,0.013452444639470843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,256,0.00282577777074443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,8192,0.219704892900255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,1536,0.014512000812424554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,16384,0.05162577827771505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,2560,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,10240,0.01556711064444648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,5120,0.008395555946562026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,3072,0.14756445089975992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,7168,0.03267911076545715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,2048,0.009696889254781935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,1536,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,3072,0.0053084443012873335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,3072,0.00906577789121204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,256,0.008288000192907121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,4096,0.007627555893527136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,8192,0.013239111337396832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,768,0.017179555363125272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,256,0.022457778453826904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,8192,0.023708444502618577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,2048,0.01963911122745938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,3584,0.0704782207806905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,6144,0.05407289001676771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,1536,0.004556444370084339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,3584,0.011786667009194693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,32,0.010002666877375709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,10240,0.02425866656833225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,1536,0.005927111125654644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,1024,0.0042257776690853965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,2048,0.006239111224810283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,512,0.016555555992656283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,2048,0.02699111070897844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,4096,0.010474666953086853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,12288,0.01959466603067186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,3072,0.023111111587948267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,128,0.006214222146405115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,65536,4.4902089436848955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,512,0.006634666687912411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,3584,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,256,0.02064533366097344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,6144,0.02294933299223582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,2048,0.02356622285313076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,4096,0.013419555293189155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,12288,0.0232631117105484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,4096,0.01668444441424476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,5120,0.011260444091426002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,64,0.006605333338181178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,3072,0.011022222538789114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,12288,0.02684533264901903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,32,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,7168,0.016388444436921012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,7168,0.19521955649058023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,768,0.01519466605451372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,3584,0.005071111023426056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,768,0.005867555737495422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,3584,0.018418666389253404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,2048,0.01184177812602785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,1536,0.006405333264006509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,7168,0.012599999705950419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,1536,0.008345777789751688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,512,0.005617777920431561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,2048,0.020617778102556866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,768,0.06453955835766263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,128,0.01239999963177575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,1024,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,10240,0.051791111628214516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,8192,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,65536,0.008978666530715095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,512,0.011800000237094032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,6144,0.0090515555606948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,3072,0.014141332772043018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,16384,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,12288,0.059254222446017794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,512,0.00942755573325687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,2048,0.3647119998931885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,10240,0.05000533329115974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,65536,0.35597509808010525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,4096,0.012231999801264869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,768,0.032050665881898664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,3584,0.011766222616036734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,128,0.0127688886390792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,65536,0.11428799894120957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,10240,0.019819556011093985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,7168,0.03417866759830051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,6144,0.019931554794311523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,4096,0.008242666721343994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,768,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,7168,0.024334222078323364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,5120,0.014662222729788886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,2560,0.016178665889634024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,3072,0.007248889240953658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,4096,0.009023110899660323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,6144,0.017224889662530687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,2048,0.02186222208870782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,4096,0.011760888828171624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,7168,0.010668444136778513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,12288,0.017073776986863878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,12288,0.06062044699986776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,4096,0.01519733336236742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,12288,0.047797332207361855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,6144,0.009053332938088311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,1024,0.00564533347884814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,512,0.004615111069546806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,128,0.0028088889602157804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,128,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,128,0.014797333214018079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,7168,0.03669688767857022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,2048,0.18511644999186197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,7168,0.027419555518362258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,32,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,1536,0.056735999054378934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,3072,0.007258666886223688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,65536,0.07358844412697686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,3584,0.0074471111098925276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,8192,0.03424711028734843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,4096,0.01293955577744378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,128,0.002830222249031067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,16384,0.02459822263982561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,6144,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,128,0.0018684443914228016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,5120,0.01034755590889189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,12288,0.15327555603451198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,10240,0.018550222118695576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,3584,0.02776799930466546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,5120,0.008348444269763099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,8192,0.016745777593718637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,65536,0.09857244624031915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,768,0.013472889032628802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,8192,0.022096888886557683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,3072,0.007305777735180325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,65536,0.24172533882988823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,6144,0.08129422532187568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,12288,0.018548443913459778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,16384,0.06232533189985487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,3072,0.09238044420878093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,32,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,4096,0.011003555523024665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,7168,0.014503111441930136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,10240,0.0258879992696974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,512,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,7168,0.026736888620588515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,7168,0.023734221855799358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,6144,0.009007111191749573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,6144,0.04962133367856344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,768,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,3584,0.008094222181373173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,1024,0.009687110781669617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,3584,0.026053332620196875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,8192,0.012664000193277994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,2560,0.006945778098371293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,5120,0.004555555681387584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,2560,0.006958222223652734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,2560,0.012472000386979846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,128,0.013058666553762225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,2048,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,5120,0.015275556180212231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,1536,0.023031110564867657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,3072,0.007256888680987888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,16384,0.5472702450222439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,2560,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,7168,0.07716533210542467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,4096,0.007821333077218797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,1024,0.0032524443748924467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,2048,0.013414222333166333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,3072,0.007479999628331926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,6144,0.012445333103338877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,8192,0.20423821608225504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,1536,0.03119022316402859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,16384,0.028502222564485338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,10240,0.06398755311965942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,8192,0.2687671184539795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,6144,0.016125332978036668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,2048,0.014480888843536377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,1536,0.022687110635969374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,1024,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,6144,0.16974755128224692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,6144,0.02058577703105079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,3584,0.0119982221060329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,3584,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,3584,0.05336888962321811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,2048,0.011895111037625207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,32,0.0018399999373488957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,65536,0.052404443422953285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,8192,0.019701333509551156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,7168,0.010009778042634329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,65536,0.07457066906823052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,10240,0.02397777802414364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,1536,0.0035537779331207275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,8192,0.01664888858795166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,1536,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,1024,0.00795555528667238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,1536,0.028351111544503108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,1024,0.6383448706732856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,2560,0.1424648894204034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,128,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,512,0.009005332986513773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,2560,0.011067555182509951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,10240,0.01629155543115404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,768,0.011404444773991903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,3072,0.06995377937952678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,6144,0.01351288871632682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,512,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,5120,0.012086222569147745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,768,0.017539555827776592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,65536,0.9460426966349283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,10240,0.016171556380059984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,3072,0.007156444092591603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,3584,0.013410666750537025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,128,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,3584,0.011054221954610614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,3584,0.007335111498832703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,12288,0.03965333435270522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,512,0.025039111574490864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,512,0.012091555529170565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,512,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,10240,0.053808887799580894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,3072,0.0075911109646161394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,65536,0.06038844585418701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,3072,0.01587199999226464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,12288,0.019127999742825825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,512,0.009011555876996782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,5120,0.03194933467441135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,10240,0.16805599795447457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,512,0.01815466748343574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,512,0.00656444455186526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,4096,0.016584000653690763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,12288,0.040354665782716535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,4096,0.03121333320935567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,5120,0.05006666647063362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,2560,0.012981333666377597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,16384,0.3887360095977783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,2560,0.0125591109196345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,128,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,3584,0.01642577846844991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,65536,0.0705475542280409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,16384,0.017101332545280457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,1024,0.014262222581439547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,512,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,3072,0.014409777190950183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,5120,0.012423111332787408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,6144,0.018780445059140522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,5120,0.009940444595283931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,4096,0.010615111225181155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,1536,0.008651555412345463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,2048,0.007313778003056844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,512,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,4096,0.015444444285498725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,2560,0.05088000165091621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,8192,0.05470755696296692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,7168,0.020243555307388306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,768,0.015192000402344598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,512,0.006251555350091722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,12288,0.028916445043351915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,768,0.005388444496525659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,3072,0.034607112407684326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,65536,0.009320889082219865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,7168,0.012720888687504662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,16384,0.03533422284656101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,6144,0.03384533193376329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,5120,0.008286221987671321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,65536,0.07633511225382487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,6144,0.01957155598534478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,10240,0.032030221488740705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,1536,0.013059555656380124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,65536,0.26588622728983563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,4096,0.02219644520017836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,3072,0.014089778065681458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,16384,0.10416266653272842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,768,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,10240,0.10118489132987128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,2560,0.020982222424613103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,256,0.004923555586073133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,256,0.016198222835858662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,3584,0.006592000110281839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,16384,0.010905777414639791
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,32,0.002846222163902389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,6144,0.03178222311867608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,2560,0.022054221895005967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,128,0.01959822244114346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,128,0.004916444420814514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,512,0.003580444388919406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,128,0.007639110916190677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,8192,0.03105955653720432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,768,0.010018666585286459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,3072,0.006968888971540663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,16384,0.03388533327314589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,512,0.055431110991372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,12288,0.029778665966457788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,10240,0.013552889227867126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,256,0.00249422217408816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,256,0.00527555536892679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,6144,0.009366222553782994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,4096,0.006640888750553131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,768,0.008375999828179678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,256,0.0023991111665964127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,3584,0.004968889057636261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,3584,0.007985778152942657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,65536,0.04051466782887777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,7168,0.021512000097168818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,2560,0.0069066666894488865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,32,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,768,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,256,0.01792533364560869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,512,0.006569777925809224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,12288,0.01961777773168352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,7168,0.01481066644191742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,65536,0.07526044713126288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,512,0.005604444278611078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,768,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,5120,0.01650399963061015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,64,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,5120,0.00830577810605367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,10240,0.012463111016485425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,5120,0.023736889163653057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,5120,0.021563554803530376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,2048,0.009453333086437648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,16384,0.024131556351979572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,10240,0.03695288962788052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,2048,0.00796977761718962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,8192,0.22494488292270234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,2560,0.006695111178689533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,1536,0.029698666599061754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,6144,0.009680888719028896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,5120,0.009288888838556077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,128,0.01886755559179518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,4096,0.03280622098180983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,8192,0.0889973309304979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,2048,0.03372622198528714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,768,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,3584,0.007985778152942657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,5120,0.18780888451470268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,768,0.002826666666401757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,1536,0.011023110813564725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,4096,0.036200887627071805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,512,0.0025191110455327565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,3584,0.02126666737927331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,64,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,512,0.016915554801623028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,6144,0.009062222308582729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,3584,0.007844444778230455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,10240,0.02309333284695943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,768,0.014819555812411837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,2048,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,6144,0.053631110323799976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,4096,0.03294488787651062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,16384,0.05271733469433255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,32,0.0034053333931499054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,8192,0.056136886278788246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,8192,0.013453333742088742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,1024,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,4096,0.02021422154373593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,16384,0.23287823465135363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,3072,0.007322666545708974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,4096,0.01550044450494978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,10240,0.023815110325813293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,5120,0.023017777336968318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,2560,0.01037333326207267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,512,0.005919999960396025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,7168,0.02882933285501268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,128,0.005003555367390315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,512,0.015011555618709989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,2560,0.006625777731339137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,512,0.013425778183672162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,512,0.004601777841647466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,128,0.005667555663320754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,12288,0.017601778109868366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,7168,0.15033333831363255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,3584,0.015538665983412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,128,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,2560,0.008852444589138031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,2560,0.026850667264726426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,1024,0.0032008888406885993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,10240,0.023152000374264185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,1024,0.0028853333658642242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,6144,0.023582221733199224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,768,0.02074399921629164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,10240,0.017176000608338248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,4096,0.01146488885084788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,3584,0.007919111185603673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,3072,0.004243555582231945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,128,0.005625777774386936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,256,0.00519466648499171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,2048,0.015104000767072042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,256,0.006640888750553131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,32,0.006904888898134232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,512,0.01651733285850949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,7168,0.030439999368455675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,3584,0.021872000561820135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,512,0.002508444504605399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,1536,0.0038746665749284956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,2560,0.007010666860474481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,2560,0.028888887829250757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,1536,0.009209777745935652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,128,0.006272000157170826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,768,0.003882666842805015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,12288,0.05489244394832187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,32,0.005624889085690181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,1536,0.022272000710169475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,3584,0.011037333144081963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,2048,0.009668444593747457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,16384,0.02607822252644433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,6144,1.0592720243665907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,6144,0.012072000238630505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,65536,0.09619200229644775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,1536,0.010795555180973478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,65536,0.0841111143430074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,65536,0.07713777489132352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,4096,0.012368888490729861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,6144,0.009240888886981541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,65536,0.13934489091237387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,5120,0.015135111080275642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,8192,0.022106667359670002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,512,0.006627555522653792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,3072,0.005617777920431561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,128,0.006026666611433029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,128,0.0021768889079491296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,3072,0.02883022361331516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,16384,0.32781955930921763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,3584,0.018246221873495314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,12288,0.2287022272745768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,256,0.015127110812399121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,7168,0.009639999932712978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,10240,0.0414631101820204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,8192,0.013081777426931592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,8192,0.038400888442993164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,10240,0.008000888758235509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,12288,0.22863377465142143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,1024,0.021278222401936848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,10240,0.012089778151777057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,64,0.0021760000122918025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,1024,0.013423110875818463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,8192,0.09328177902433608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,65536,0.06735289096832275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,3072,0.13718310991923013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,16384,1.0259786181979709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,1024,0.023964444796244305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,2048,0.00795288880666097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,10240,0.15902133782704672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,4096,0.007632888853549957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,3072,0.007272889216740926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,65536,0.22205777963002524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,1536,0.006270222365856171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,5120,0.014084445105658637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,128,0.0022088889446523455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,128,0.003543111185232798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,10240,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,10240,0.018234666850831773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,10240,0.01651022169325087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,128,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,128,0.005538666827811136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,3072,0.004767111192146937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,7168,0.08870489067501491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,256,0.002471999989615546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,1024,0.012840888566441007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,8192,0.06102310948901706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,512,0.011038222246699862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,2048,0.006203555398517185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,16384,0.02366222275627984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,256,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,12288,0.03536355495452881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,1536,0.018222222725550335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,3584,0.017210667332013447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,3072,0.025681777132882014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,10240,0.021187555458810594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,256,0.0041928887367248535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,128,0.002879111096262932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,3072,0.007246221933099959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,768,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,3072,0.01753866672515869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,6144,0.009316444396972656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,128,0.014133334159851074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,6144,0.05487644341256884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,10240,0.015160888433456421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,128,0.019366222951147292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,16384,0.04522755410936144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,16384,0.017501332693629794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,512,0.05365066727002462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,128,0.01753777762254079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,16384,0.034443553951051496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,3584,0.012748444245921241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,3072,0.010096000300513374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,2048,0.021942221456103857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,1536,0.006238222122192383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,256,0.002472888885272874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,6144,0.01624000072479248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,10240,0.029079112741682265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,1024,0.009373333719041612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,12288,0.018575110369258456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,6144,0.017775999175177682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,2048,0.006588444527652528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,128,0.0020684444655974707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,6144,0.02219999995496538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,32,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,3072,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,16384,0.2650177743699816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,8192,0.1507155497868856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,4096,0.00848444468445248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,128,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,256,0.015125332607163323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,512,0.015084443820847405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,2048,0.008788444101810455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,12288,0.032278223170174494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,8192,0.014476444986131458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,256,0.012108444339699216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,16384,0.07437955670886569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,8192,0.011988444460762871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,16384,0.029776000314288672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,3072,0.03161333335770501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,1536,0.00871288859181934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,2560,0.014089778065681458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,12288,0.02273155583275689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,7168,0.025768889321221247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,5120,0.008708444734414419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,4096,0.007617777420414819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,12288,0.014135999812020196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,8192,0.024880000286632117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,5120,0.01648622254530589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,64,0.0018497777895794974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,2560,0.07714044385486178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,128,0.002116444417172008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,1536,0.013749333719412485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,65536,0.09151377942827012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,5120,0.008419555922349295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,8192,0.0063742221229606206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,4096,0.02830844455295139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,256,0.005532444351249271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,768,0.0166311115026474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,16384,0.030377778742048476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,65536,0.10222755538092719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,1024,0.0668355557653639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,16384,0.05728533532884386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,3072,0.029696000946892634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,128,0.004553777890072929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,10240,0.018904000520706177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,5120,0.02540622154871623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,8192,0.03163822160826789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,3072,0.01755199995305803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,8192,0.025348444779713947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,16384,0.0403511126836141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,3072,0.014071111877759298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,1536,0.009214222431182861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,2048,0.05907377931806776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,256,0.007667555577225155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,12288,0.5095413525899252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,768,0.02902311086654663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,6144,0.010821333361996545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,5120,0.07407733466890123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,65536,0.15317600303226048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,65536,1.4749529096815321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,128,0.0075937774446275495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,5120,0.011901333100265928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,3584,0.013960889644092984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,3584,0.007629333270920648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,65536,0.123943116929796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,768,0.0032044444233179092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,65536,0.0799600018395318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,10240,0.015159999330838522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,65536,0.06054844458897909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,256,0.0063368889192740125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,128,0.004984000076850255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,16384,0.020240889655219186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,12288,0.02201244400607215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,4096,0.011737777955002256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,2048,0.014141332772043018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,3072,0.023403555154800415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,1024,0.005981333139869902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,2560,0.023176888624827068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,6144,0.004219555606444676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,768,0.005684444473849402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,65536,0.07527911000781588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,64,0.007961778177155389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,512,0.002542222125662698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,1024,0.0073902226156658595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,128,0.002512000087234709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,3584,0.029490666257010564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,5120,0.010668444136778513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,8192,0.023423999547958374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,768,0.003555555724435382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,8192,0.01922044489118788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,3584,0.02474844455718994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,16384,0.04799911048677233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,32,0.0021955555097924336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,10240,0.013805333111021252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,1536,0.011872889267073737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,7168,0.009850666754775578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,3072,0.02741777731312646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,256,0.002509333400262727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,65536,0.07592444287406074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,6144,0.034295111894607544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,3072,0.027269333600997925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,12288,0.02370133333735996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,768,0.02825066778394911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,10240,0.01663555536005232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,768,0.017177777157889474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,2048,0.011016888750924004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,512,0.005209777918126848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,10240,0.057625777191585965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,2048,0.012048888537618848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,1536,0.011433777709801992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,6144,0.022974222898483276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,512,0.00831911133395301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,3072,0.015496888094478183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,1024,0.0032675556010670135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,8192,0.023578666978412207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,10240,0.018602665927675035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,65536,0.027934223413467407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,3072,0.014729777971903482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,2560,0.015807999504937064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,3072,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,6144,0.0103493332862854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,8192,0.01961955593691932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,16384,0.10392266511917114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,2048,0.01385244478782018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,8192,0.15051200654771593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,1024,0.011797332929240333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,12288,3.9850311279296875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,512,0.06453333298365276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,6144,0.036575112077924944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,10240,0.03307377629809909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,4096,0.011553777588738335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,65536,0.09033688571718003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,16384,0.03570133447647095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,65536,0.08561866813235813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,65536,0.07995111412472196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,12288,0.01886133352915446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,12288,0.020019556085268658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,128,0.01353422221210268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,3072,0.03559466534190708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,1536,0.046720888879564076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,3072,0.1293262243270874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,2048,0.06260177824232313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,2560,0.027572444743580286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,2048,0.019997333486874897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,6144,0.016903110676341586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,2048,0.006030222194062338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,65536,0.11387022336324056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,128,0.004872888740566042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,256,0.003546666767862108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,512,0.01706044375896454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,7168,0.10923733313878377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,512,0.006625777731339137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,768,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,6144,0.026183999247021143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,3584,0.014094221923086377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,1024,0.011042666931947073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,6144,0.007056000332037608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,1024,0.013384888569513956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,2560,0.023373333944214716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,3072,0.008022222254011366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,7168,0.017246221502621967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,1024,0.021240888370407954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,4096,0.11049422290590073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,512,0.005978666659858491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,2048,0.004648000001907349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,7168,0.014398222168286642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,2048,0.01921244462331136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,512,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,128,0.014141332772043018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,2560,0.016881777180565726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,12288,0.014073777529928418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,3584,0.07565244701173571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,8192,0.017868444323539734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,8192,0.01071111112833023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,4096,0.2356142203013102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,2560,0.022983110613293115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,12288,0.047183109654320605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,6144,0.029707554313871596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,512,0.01963911122745938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,6144,0.022634666826989915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,8192,0.010378667049937779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,65536,0.08864621983634101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,3584,0.03272799981964959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,16384,0.04191822144720289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,12288,0.020474667350451153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,8192,0.08261599805619982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,1536,0.04138844543033176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,2560,0.024938666158252295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,512,0.014992000328169929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,1024,0.020301333732075162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,512,0.011316444310877057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,3072,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,16384,0.0191920002301534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,12288,0.40816712379455566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,256,0.007265778051482306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,5120,0.03608444333076477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,2560,0.014391111003028022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,16384,0.011062222222487131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,256,0.003645333151022593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,16384,0.07581866449779935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,128,0.004928888960017098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,32,0.0018648888087934917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,5120,0.008311111066076491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,2560,0.03197333216667175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,64,0.002879111096262932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,128,0.0028373334142896864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,65536,0.04416444566514757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,7168,0.018541332748201158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,3584,0.007269333634111617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,1536,0.003527111063400904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,64,0.0018871112002266778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,64,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,128,0.012720000412729053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,4096,0.10761688815222846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,8192,0.007120889094140794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,128,0.00904088881280687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,256,0.016501333978441026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,256,0.0158595558669832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,6144,0.021551110678248938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,6144,0.009009777671760982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,128,0.02460177739461263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,6144,0.009830222361617619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,3072,0.026351110802756414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,7168,0.037568890386157565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,5120,0.01257155587275823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,1536,0.005592000153329637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,16384,0.01682844426896837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,32,0.030729777283138696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,8192,0.1857395569483439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,1024,0.0032239999208185407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,5120,0.02644799980852339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,4096,0.025770665870772466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,2048,0.006219555520349079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,8192,0.0158924443854226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,3072,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,256,0.016184889607959323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,3584,0.08609866433673435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,2560,0.004836444639497333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,8192,0.041942222250832446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,1024,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,256,0.3199662102593316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,768,0.005618666609128316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,128,0.011366222467687396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,5120,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,5120,0.013087111214796701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,2560,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,64,0.0355057782597012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,256,0.004899555610285865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,128,0.025432889660199482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,7168,0.01860000027550591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,8192,0.13781866762373182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,65536,0.28686221440633136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,65536,0.16567822297414145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,128,0.012055111428101858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,6144,0.012061333490742577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,65536,0.06543199883566962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,2048,0.04134933484925164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,16384,0.020414221617910597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,2048,0.012303111453851065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,5120,0.017846221725145977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,128,0.00519466648499171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,3072,0.00701422244310379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,12288,0.0321706665886773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,12288,0.005683555371231503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,1536,0.008056888977686564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,1024,0.0029288888391521242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,1536,0.015085332923465304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,256,0.0020471111767821843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,3072,0.019321777754359774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,12288,0.009683555199040307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,16384,0.027262222435739305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,2560,0.0074026667409473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,512,0.03680888811747233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,16384,0.01204533295498954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,4096,0.011714666253990598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,768,0.005978666659858491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,3584,0.012810666528013019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,4096,0.01686755485004849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,1024,0.005997333261701796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,256,0.019370666808552213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,1024,0.01276266657643848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,2560,0.009656888743241629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,1024,0.003133333391613431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,4096,0.04657155606481764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,128,0.016506666938463848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,128,0.011743110915025076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,6144,0.04135555691189236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,512,0.00253866674999396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,128,0.01479555500878228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,65536,0.34286755985683864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,16384,0.025641777449183997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,10240,0.043084445926878184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,6144,0.04387199878692627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,32,0.005678222411208683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,512,0.011414222419261932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,2048,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,512,0.0031840000301599503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,512,0.00794222205877304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,10240,0.013111111190583972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,512,0.01714666684468587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,3584,0.007965332931942409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,256,0.003452444449067116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,512,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,10240,0.012392000191741519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,3072,0.008674666285514832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,4096,0.04325955443912082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,16384,0.05478755633036295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,64,0.00563822231358952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,12288,0.05482666691144308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,2048,0.013046222428480784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,256,0.014488000008794995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,768,0.005955555372767978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,256,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,768,0.006257777826653586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,5120,0.029096000724368628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,768,0.006029333505365584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,65536,0.105804443359375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,12288,0.262388441297743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,2560,0.010060444474220276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,512,0.002891555635465516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,12288,0.02082577844460805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,3072,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,7168,0.012368000215954252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,3072,0.021612443857722815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,768,0.02239022155602773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,2048,0.01648533344268799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,8192,0.0613973339398702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,1024,0.03484977616204156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,6144,0.020597333709398907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,2048,0.006227555374304454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,2560,0.008664000365469191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,3584,0.04431822233729892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,1536,0.00351555562681622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,6144,0.01131199962562985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,65536,0.20594844553205702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,3072,0.01483022173245748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,768,0.025419554776615564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,12288,0.01960711181163788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,10240,0.019313777486483257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,128,0.004866666677925321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,256,0.012364444633324942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,2560,0.006586666736337874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,10240,0.1508302158779568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,65536,0.1382008923424615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,5120,0.010464000205198923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,2048,0.017665778597195942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,7168,0.02421422302722931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,512,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,2560,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,1024,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,3072,0.005315555466545953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,16384,0.011169777976142036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,768,0.007644444704055786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,768,0.007988444632954067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,2560,0.09878400299284194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,8192,0.012844444149070315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,8192,0.01847199930085076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,1024,0.0034826666944556763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,2048,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,8192,0.02937155630853441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,8192,0.039410667286978826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,256,0.014103111293580798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,7168,0.015480000111791821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,4096,0.015481778317027621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,64,0.007311999797821045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,16384,0.49755112330118817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,8192,0.027756444282001917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,4096,0.009641778137948778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,4096,0.03943555553754171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,2048,0.012704888979593912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,12288,0.019380443625979953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,2048,0.011032000184059143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,512,0.0052151112920708126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,1536,0.0677200025982327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,64,0.005052444421582752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,6144,0.010504888991514841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,4096,0.8202915721469455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,512,0.01074488874938753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,3584,0.05589333507749769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,3584,0.01590755581855774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,10240,0.12714844279819065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,1024,0.012114666402339935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,3072,0.006251555350091722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,12288,0.01835644410716163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,6144,0.009319110876984065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,7168,0.014125333891974555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,5120,0.008912888665994009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,4096,0.029935999049080744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,1536,0.12004889382256402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,32,0.008058666355080074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,8192,0.01254666679435306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,2048,0.004663111021121343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,512,0.0052586665583981415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,5120,0.010775999890433418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,5120,0.008612444831265343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,7168,0.0188711103465822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,3584,0.015488000379668342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,64,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,1536,0.01594577729701996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,12288,0.349942233827379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,8192,0.004865777575307422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,1536,0.0035413333939181436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,1536,0.00658044425977601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,6144,0.0863386657502916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,6144,0.009010666774378883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,8192,0.014775110615624322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,6144,0.013773333695199756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,1024,0.007323555648326874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,8192,0.01109955542617374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,32,0.0018533333722088074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,64,0.025568000144428674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,65536,0.17254311508602568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,2048,0.006437333507670297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,2048,0.009296889106432596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,2560,0.01256977766752243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,16384,0.11527821752760146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,2048,0.0041555555330382455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,1024,0.005879111174080107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,8192,0.38050399886237246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,512,0.005592000153329637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,8192,0.012621333201726278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,2048,0.12509332762824163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,1024,0.005576000031497743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,1024,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,10240,0.18513422542148164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,2048,0.02529866662290361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,2048,0.04786844385994805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,256,0.01034044474363327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,32,0.002480888946188821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,1024,0.008743111458089616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,128,0.002509333400262727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,7168,0.012467555701732635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,256,0.013796444568369122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,65536,1.6394915050930445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,256,0.010436444646782346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,10240,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,6144,0.013561777770519257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,7168,0.014774221513006421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,32,0.005952888892756567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,10240,0.020026667250527274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,7168,0.05376177695062426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,10240,0.06817689206865099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,64,0.0053297777970631914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,16384,0.0236462222205268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,512,0.0052586665583981415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,64,0.0022435554613669715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,12288,0.01536711057027181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,16384,5.197628445095487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,16384,0.055600000752343066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,4096,0.09814044502046372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,768,0.018568888306617737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,5120,0.03603822324011061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,64,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,3584,0.023677332533730402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,768,0.003912000192536248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,2048,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,12288,3.8341590033637156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,8192,0.01204177737236023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,16384,0.02665600015057458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,65536,4.158410814073351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,512,0.014469333820872836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,256,0.016170667277442086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,5120,0.011383111278216044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,512,0.005978666659858491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,4096,0.008627555436558193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,65536,0.12201510535346137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,16384,0.5163760185241699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,512,0.0031928889867332247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,6144,0.01756888959142897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,1024,0.007096000015735626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,6144,0.027385777897304956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,3584,0.009933333430025313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,5120,0.04195377892918057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,6144,0.1127199994193183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,2560,0.008357333640257517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,4096,0.016806221670574613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,1024,0.02789333462715149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,128,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,2048,0.006680000159475539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,2560,0.180550217628479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,4096,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,1536,0.0038702223036024305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,65536,0.06247199906243218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,1024,0.005960888746711943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,2048,0.006614222294754452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,5120,0.008649778034951951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,7168,0.010191111101044549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,4096,0.007598222129874759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,6144,0.009367999931176504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,12288,0.15841955608791775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,65536,0.06630311409632365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,10240,0.020633776982625324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,256,0.011412444214026133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,6144,0.033019555939568415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,3584,0.028386665715111628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,3584,0.023344000180562336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,10240,0.09116799963845147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,65536,0.07567022244135539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,2048,0.014153778553009033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,16384,0.04195911023351881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,4096,0.014367999302016364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,128,0.014840889308187695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,128,0.00226488895714283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,6144,0.04759822289148966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,3072,0.009399111072222391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,2560,0.03045155604680379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,512,0.006643555644485686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,5120,0.008314666648705801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,8192,0.15965421994527182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,10240,0.016975111431545682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,2560,0.006968888971540663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,10240,0.016705777910020616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,128,0.03822222352027893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,64,0.0021395554973019496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,512,0.0025351111673646504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,5120,0.007601777712504069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,4096,0.009003555609120263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,512,0.002834666727317704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,8192,0.03231199913554721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,2048,0.0042204442951414315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,12288,0.3522968822055393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,1024,0.0058204444746176405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,2048,0.006937777830494775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,5120,0.00868088917599784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,2048,0.014045332868893942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,2048,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,7168,0.016154666741689045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,3072,0.0073315559162033936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,5120,0.016047999262809753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,128,0.002171555534005165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,128,0.01440355512830946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,65536,0.1472355524698893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,5120,0.022688888841205176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,3584,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,3072,0.010054222411579555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,768,0.018602665927675035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,3072,0.019679110911157396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,65536,0.9240978029039172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,1024,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,6144,0.14221955670250788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,65536,0.07755911350250244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,768,0.014110222458839417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,64,0.00184888889392217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,3584,0.01482222146458096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,768,0.005177777674463061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,7168,0.017500445246696472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,3584,0.030846221579445735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,1536,0.010326222413116032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,12288,0.1000444425476922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,12288,0.023984000086784363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,256,0.004906666775544484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,2560,0.008051555189821456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,256,0.003165333428316646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,8192,0.08164355489942762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,2048,0.03869244456291199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,3584,0.007266666326257918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,1536,0.01458044515715705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,8192,0.0105262224872907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,10240,0.00777511133088006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,1536,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,12288,0.019155555301242404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,65536,0.10228088829252456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,1024,0.005882666756709416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,2048,0.006254222244024277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,768,0.00832444429397583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,3072,0.009299555586444007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,4096,0.003955555458863576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,3584,0.015912888778580558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,65536,0.07571911149554782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,10240,0.018381334013409086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,5120,0.03466044531928168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,256,0.005071111023426056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,512,0.005407111098368962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,8192,0.010703110860453712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,768,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,12288,0.5769759813944498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,64,0.00562755556570159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,2560,0.007612444460391998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,256,0.014863999353514777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,7168,0.009363555245929295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,256,0.030857778257793848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,10240,0.014592889282438489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,256,0.031975110371907554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,512,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,3584,0.03358133302794562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,7168,0.12746666537390813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,3072,0.007511999871995714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,6144,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,64,0.002583111118939188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,8192,0.049714667929543384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,65536,0.0770311090681288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,8192,0.014817777607176038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,3584,0.02854222059249878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,2560,0.02335111134582096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,2560,0.013803555733627744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,8192,0.045774221420288086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,16384,0.030202666918436687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,512,0.005210666606823604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,1024,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,128,0.022299556268586054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,5120,0.008664000365469191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,7168,0.01311822235584259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,128,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,2048,0.006918222126033571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,16384,0.019113777412308585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,1536,0.003943110919660992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,3072,0.0233262214395735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,1536,0.0035537779331207275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,1536,0.015559110376569959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,7168,0.011086222198274402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,768,0.006292444550328785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,512,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,6144,0.01752800080511305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,512,0.0024826667375034755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,2048,0.03192355566554599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,768,0.01479644411140018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,256,0.005882666756709416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,65536,0.11366311709086101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,3072,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,4096,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,4096,0.03489689032236735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,5120,0.026022222306993272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,768,0.020243555307388306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,5120,0.01479288935661316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,2560,0.00692622239391009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,128,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,7168,0.01479200025399526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,2048,0.012094222009181976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,12288,0.07546755340364245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,2560,0.052242666482925415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,768,0.010429333481523724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,128,0.006106666806671355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,7168,0.035572446054882474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,3584,0.009383111364311641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,1536,0.006918222126033571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,2048,0.025535111625989277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,16384,0.15439644124772814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,7168,0.05318489008479648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,3584,0.007293333609898885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,4096,0.01741422216097514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,16384,0.01684711045689053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,2048,0.006167111297448476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,3072,0.010095111197895473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,4096,0.007972444097201029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,65536,0.07800266477796766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,16384,0.028005331754684448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,256,0.015858666764365304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,512,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,2048,0.009291555318567488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,16384,0.023430221610599097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,12288,0.04164088765780131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,32,0.16219822565714517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,8192,0.01648266613483429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,1024,0.03299999899334378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,64,0.00795022232664956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,16384,0.09851200050777859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,2048,0.004864000197913912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,12288,0.1765520042843289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,4096,0.010510221951537661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,16384,0.2821893427107069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,128,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,64,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,256,0.004956444518433677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,7168,0.009445333646403419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,7168,0.010178666975763109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,3584,0.010699555277824402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,32,0.0022408889813555614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,16384,0.0616862244076199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,7168,0.012456888953844706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,1024,0.005666666560702854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,128,0.0053173332578606075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,256,0.014476444986131458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,2560,0.02068444424205356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,7168,0.013836444252067141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,512,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,3072,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,128,0.01202488856183158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,16384,0.14853421847025552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,5120,0.030645334058337744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,7168,0.013781333135233985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,32,0.0038480001191298165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,12288,0.05374933282534281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,10240,0.042259554068247475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,1024,0.004184000194072723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,256,0.013795555465751223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,768,0.01443377799457974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,65536,3.147484461466471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,2048,0.02942044536272685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,512,0.0029066666546795103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,2560,0.011006222003036074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,7168,0.013723555538389417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,8192,0.013540444274743399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,128,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,768,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,5120,0.09715733263227676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,4096,0.09082666370603774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,6144,0.02295822236273024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,768,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,8192,0.01401600076092614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,7168,0.03274666600757175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,65536,0.04727910955746969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,4096,0.03137866655985514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,12288,0.01774488886197408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,512,0.002536888958679305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,512,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,5120,0.01129422254032559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,10240,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,4096,0.013215111361609565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,12288,0.0195777780479855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,10240,0.019944889677895438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,5120,0.03840800126393636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,5120,0.009303111169073317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,3072,0.04526755544874403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,2048,0.025798221429189045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,6144,0.011471110913488599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,768,0.0035048888789282907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,7168,0.011930666863918304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,1024,0.005628444254398346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,128,0.0021048889805873237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,7168,0.08077777756585015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,128,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,2048,0.0236151119073232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,3072,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,10240,0.0547653337319692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,6144,0.008987555901209513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,10240,0.014200000299347771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,65536,0.12209244569142659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,3072,0.01681688924630483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,64,0.020915556285116408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,64,0.0018177777528762817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,65536,0.08183377981185913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,4096,0.020938666330443487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,256,0.029666665527555678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,3072,0.030481778913074072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,512,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,12288,0.23556711938646105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,1536,0.006232888748248418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,12288,0.013711111413107978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,2048,0.005127999931573868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,5120,0.013080000049538083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,512,0.015858666764365304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,10240,0.02093688812520769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,7168,0.46762577692667645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,32,0.004214222232500712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,1536,0.010072000324726105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,7168,0.007592889169851939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,6144,0.014298665854665967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,8192,0.010951999988820819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,768,0.012455110748608908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,1536,0.005942222144868638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,10240,0.02609955602222019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,8192,0.013445333474212222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,3072,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,1024,0.04489333430926005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,64,0.008298666940795051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,4096,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,12288,0.018795554836591084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,6144,0.010957332948843638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,4096,0.02899644441074795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,65536,0.058134224679734975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,12288,0.1263048913743761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,3584,0.007290666302045186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,256,0.0021991110924217436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,128,0.015830222103330824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,256,0.005214222189452913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,2560,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,5120,0.008629333641793992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,512,0.014045332868893942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,3072,0.00692622239391009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,16384,0.020907556017239887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,1536,0.013730666703648038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,10240,0.014142221874660917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,2048,0.005933333188295364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,2560,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,32,0.006590222318967183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,16384,0.10299111074871486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,1536,0.5836515426635742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,1024,0.006950221955776215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,65536,0.011223999990357293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,2560,0.0350524451997545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,16384,0.013051555388503604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,16384,0.07223466369840834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,1024,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,16384,0.022989332675933838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,12288,0.018201778332392376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,65536,0.18693778249952528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,512,0.005587555468082428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,64,0.002143111079931259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,1024,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,8192,0.036291556225882635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,10240,0.031119111511442397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,3072,0.007613333563009898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,2048,0.004938666605287128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,2560,0.08435377809736465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,16384,0.024355555574099224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,3072,0.015610666738616096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,3584,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,512,0.005604444278611078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,3072,0.033687111404207014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,1536,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,5120,0.08311733272340563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,12288,0.04403466648525662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,4096,0.014565333724021912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,768,0.01787644459141625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,128,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,65536,0.08052621947394477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,6144,0.029581334855821397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,8192,0.01789066692193349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,1024,0.0031786666562159858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,16384,0.02307555576165517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,256,0.004953777624501122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,3584,0.02630222174856398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,1536,0.010496888723638324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,12288,0.02325511144267188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,65536,0.09015822410583496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,768,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,4096,0.006962666908899943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,12288,0.4939289093017578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,16384,0.10171111424763997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,128,0.002838222309947014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,2048,0.008449777960777283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,4096,0.03451111250453525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,5120,0.04890577660666572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,12288,0.030693332354227703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,16384,0.050203555160098605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,1536,0.0035422220826148987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,2560,0.048623998959859215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,7168,0.02199111051029629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,12288,0.004905777672926585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,512,0.013471999930010902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,4096,0.004233777936961916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,128,0.011067555182509951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,3584,0.010874666273593903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,12288,0.024028443627887305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,16384,0.06556444697909884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,10240,0.016897777716318767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,65536,0.059141331248813205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,12288,0.0414008895556132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,3072,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,1024,0.0038657776183552216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,4096,0.007876444194051955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,2048,0.006236444330877728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,12288,0.019077334139082167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,16384,0.044920888211992055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,2560,0.007000889215204451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,16384,0.11204355292850071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,65536,0.06345511145061918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,6144,0.03086577852567037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,10240,0.03150044547186957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,4096,0.010810666614108615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,10240,0.012413333687517377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,256,0.018022222651375663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,3584,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,7168,0.010323555933104621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,7168,0.02828622195455763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,16384,0.019484443797005545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,1024,0.016217778126398723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,512,0.005884444548024072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,4096,0.029151999288135107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,5120,0.017679999272028606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,2048,0.006321777900060018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,8192,0.005959110955397288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,2048,0.015119999647140503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,8192,0.011076444553004371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,256,0.004987555659479565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,6144,0.014021333720948962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,7168,0.012596444123321109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,3584,0.010635555618339114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,256,0.0219217770629459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,768,0.00869777798652649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,3072,0.010225777824719747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,128,0.004874666531880696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,3584,0.03407377666897244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,65536,1.3953226937188044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,1024,0.009008000294367472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,128,0.01549600064754486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,1024,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,2048,0.019893333315849304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,1536,0.02315644423166911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,1024,0.0058542220956749385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,1024,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,3072,0.03163733416133457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,5120,0.016125332978036668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,128,0.015589333242840238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,2048,0.004570666700601578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,3072,0.007002666592597961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,8192,0.05999555852678087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,6144,0.041532443629370794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,5120,0.10172266430324978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,7168,0.07481333282258776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,7168,0.010001777774757808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,65536,0.17795911100175646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,7168,0.018385777870814007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,1536,0.006058666855096817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,5120,0.008647110727098253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,1536,0.014134221606784396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,7168,0.022924444741672937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,7168,0.05506933397716946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,7168,0.022336888644430373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,768,0.058911111619737416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,4096,0.012279111478063794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,512,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,7168,0.035178667969173856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,7168,0.07377422518200345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,2048,0.006598222172922558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,3584,0.016135111451148987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,16384,0.026029333472251892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,1536,0.009380444056457942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,7168,0.05830755498674181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,7168,0.07724266582065158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,1024,0.003165333428316646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,6144,0.00904888908068339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,5120,0.010999110837777456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,1536,0.00508355556262864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,2560,0.015792888071801927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,6144,0.02381155557102627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,128,0.013797333670987023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,256,0.016184889607959323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,6144,0.03425511055522495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,512,0.002793777734041214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,5120,0.17194488313463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,128,0.019475556082195707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,10240,0.0718640022807651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,512,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,12288,0.04523022307289971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,3072,0.01514755520555708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,3584,0.1115457746717665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,3584,0.00756444450881746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,1024,0.014864888456132678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,256,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,10240,0.32354309823777944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,128,0.007648888561460707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,256,0.016916443904240925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,3584,0.028365333875020344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,2560,0.022595556245909795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,12288,0.01828977796766493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,3584,0.012228444218635559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,768,0.016735111673672993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,3584,0.022950222094853718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,12288,0.05564800235960218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,768,0.005432888865470886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,3584,0.03453066613939073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,128,0.04945333467589485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,6144,0.009358222285906473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,12288,0.040928890307744346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,3072,0.008284444610277811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,7168,0.010371555884679159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,10240,0.010669333239396414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,128,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,3584,0.019869332512219746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,10240,0.01921866668595208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,64,0.005734222216738596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,16384,1.038977728949653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,1536,0.013626666532622443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,65536,0.16207377115885416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,2560,0.018914666440751817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,10240,0.02583822276857164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,12288,0.08899199962615967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,6144,0.07867733637491862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,12288,0.026924444569481745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,65536,0.07769511143366496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,10240,0.027071999178992376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,16384,0.01463733282354143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,1024,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,2048,0.021213332811991375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,6144,0.009671111073758867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,10240,0.044469333357281156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,3584,0.0239982224173016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,3072,0.006979555719428592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,6144,0.013205333716339536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,65536,0.07528355386522081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,64,0.009691555466916826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,7168,0.011023110813564725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,7168,0.01443733274936676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,1024,0.01955111159218682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,32,0.0038906666967603895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,6144,0.19937154981825086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,768,0.01482311056719886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,65536,0.2772604359520806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,2048,0.027264888087908428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,768,0.015192000402344598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,1536,0.0063031112982167145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,32,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,7168,0.01993600030740102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,8192,0.08400266700320774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,4096,0.012319111161761813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,128,0.0025155554629034465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,8192,0.0221742226017846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,65536,0.07445422146055433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,5120,0.008393777741326226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,16384,0.03728444377581278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,5120,0.013239111337396832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,5120,0.015106666419241162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,16384,0.0599235561158922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,3072,0.007271111011505127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,10240,0.0058746664888328975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,2048,0.004513777792453766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,3072,0.03220977716975742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,3072,0.014530667000346713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,1024,0.012880889077981314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,65536,0.020970667401949566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,5120,0.02588533361752828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,65536,0.24220445421006945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,8192,0.030495110485288832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,6144,0.03384266628159417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,65536,0.06128622425927056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,3072,0.10939733187357585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,768,0.01759022143152025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,12288,0.03429333368937174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,3072,0.024023110667864483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,512,0.0028151110228565005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,8192,0.01854222185081906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,5120,0.017232000827789307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,10240,0.016153777639071148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,768,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,1024,0.016538666354285348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,1024,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,4096,0.02439022229777442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,12288,0.04344711038801405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,3584,0.007640889121426477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,7168,0.010996444357766045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,8192,0.015787555111779105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,2560,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,12288,0.025850666893853083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,128,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,10240,0.05921510855356852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,65536,0.07123555739720662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,3072,0.012477333347002665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,768,0.006048888795905643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,7168,0.030949334303538006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,16384,0.1211377779642741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,512,0.002524444419476721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,2048,0.04626133375697666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,5120,0.030682666434182063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,1536,0.006197333335876465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,2048,0.021125333176718816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,6144,0.02299022177855174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,768,0.01700888905260298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,10240,0.032015111711290144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,65536,0.07462577687369452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,65536,0.09043111403783162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,3584,0.32570044199625653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,128,0.014480888843536377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,4096,0.018957333432303537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,4096,0.032039112514919706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,4096,0.009311111436949836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,1024,0.005643555687533484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,8192,0.01642933322323693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,32,0.005670222143332164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,1536,0.013883555928866068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,3584,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,3072,0.015537778536478678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,12288,0.019087110956509907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,7168,0.011864888999197217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,10240,0.03506222367286682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,256,0.041951109965642296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,4096,0.02737066646416982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,12288,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,128,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,32,0.006030222194062338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,256,0.005395555661784277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,3072,0.21332177850935194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,6144,0.029247111744350855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,4096,0.019684443871180218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,65536,0.07650755511389838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,128,0.012430222498046027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,5120,0.00868088917599784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,4096,0.030833777454164293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,512,0.007291555404663086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,256,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,128,0.022319111559126113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,64,0.0018542222678661346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,1024,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,32,0.0069288888739215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,1536,0.019911110401153564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,16384,0.046959112087885536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,6144,0.03616888986693488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,6144,0.3223724365234375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,65536,0.22286044226752388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,12288,0.021516443954573736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,256,0.00203822222020891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,2560,0.03194488750563727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,1024,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,65536,0.08867822090784709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,5120,0.04814133379194471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,128,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,65536,0.08459466695785522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,1536,0.003556444413132138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,12288,0.0218551109234492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,128,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,3072,0.17984267075856528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,65536,0.07563910881678264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,16384,0.046556442975997925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,3072,0.006984000404675801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,1024,0.013805333111021252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,128,0.005210666606823604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,3072,0.009350222018029954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,3584,0.023725334140989516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,1024,0.003903999924659729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,1536,0.012081777883900536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,1536,0.016061334146393668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,16384,0.02348800003528595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,8192,0.010383110907342698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,6144,0.03401244348949856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,7168,0.0963546633720398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,3584,0.007633777956167857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,768,0.016174222032229107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,12288,0.1848355531692505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,1024,0.017869333426157635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,6144,0.00665244460105896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,512,0.35941778288947207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,6144,0.009307555854320526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,6144,0.050231999821133085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,512,0.008967111508051554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,2560,0.024217777782016333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,1536,0.011295110815101199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,3584,0.028111110130945843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,512,0.005333333379692501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,128,0.002190222135848469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,7168,0.06795022222730848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,768,0.011437333292431302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,7168,0.035544001393847995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,16384,0.10803555779986912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,12288,0.023694222172101338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,8192,0.03842666745185852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,1024,0.011367110742463006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,8192,0.0173680004146364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,8192,0.050903111696243286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,512,0.006325333482689328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,128,0.0018497777895794974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,1536,0.006200888918505774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,4096,0.00794311116139094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,2560,0.007301333049933116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,10240,0.01575644479857551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,8192,0.022282666630215112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,5120,0.012095111111799875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,3584,0.03743288914362589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,128,0.05960710843404134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,8192,0.014835556348164877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,7168,0.018936889039145578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,768,0.0032266666077905228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,6144,0.02027555637889438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,1024,0.004643555730581284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,1024,0.02538577715555827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,12288,0.020576889316240948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,3072,0.03725422090954251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,3584,0.030136889881557886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,6144,0.01755466726091173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,65536,0.06895377900865343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,16384,0.017100445098347135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,12288,0.017535999417304993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,4096,0.009005332986513773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,16384,0.01477688882086012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,3584,0.06671555836995442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,2560,0.022646221849653456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,512,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,6144,0.019729778170585632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,1024,0.00590844452381134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,3072,0.007258666886223688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,64,0.04532000091340807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,12288,0.021009777983029682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,3072,0.010976889067225985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,4096,0.016247999336984422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,2560,0.07500088877148099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,768,0.06132533152898153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,16384,0.026331555512216356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,12288,0.025342222717073228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,8192,0.010439111126793755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,3584,0.007974222302436829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,7168,0.12383911344740127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,2560,0.006554666492674086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,16384,0.02535999980237749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,3584,0.00867555538813273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,768,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,32,0.00573955559068256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,2048,0.07138755586412218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,1536,0.006230222268237009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,16384,0.031179553932613794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,12288,0.03774044579929776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,1536,0.007080888582600488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,1024,0.008996444443861643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,3072,0.024979554944568213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,16384,0.027257778578334387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,10240,0.2573173311021593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,5120,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,3072,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,256,0.002480888946188821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,64,0.00351555562681622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,12288,0.04749155706829495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,128,0.01481599940194024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,10240,0.18682400385538736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,32,0.00796622203456031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,8192,0.02169955604606205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,256,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,65536,0.2055048942565918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,4096,0.025024889243973628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,10240,0.015600000818570455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,768,0.008363555702898238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,256,0.006401777681377199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,768,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,8192,0.015620445211728414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,1536,0.005982222242487802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,256,0.0025253333151340485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,4096,0.19653956095377603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,12288,0.02776355544726054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,8192,0.008367111285527548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,6144,0.01700444519519806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,768,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,3072,0.007140444384680853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,65536,0.1641075611114502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,65536,0.3842311170366075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,16384,0.03122577733463711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,768,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,3584,0.015767110718621146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,1024,0.003920888735188378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,10240,0.022671111755900916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,5120,0.10630222161610921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,2048,0.006597333484225803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,7168,0.02308799988693661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,768,0.003504000190231535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,128,0.0022311111291249595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,256,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,10240,0.034276445706685386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,65536,0.07997155851787992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,256,0.002833777831660377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,1024,0.06243289179272122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,7168,0.019476445184813607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,2048,0.006585777633719974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,256,0.002143111079931259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,1536,0.014456889695591398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,128,0.014136888914638095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,1024,0.0069191112286514705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,8192,0.0314408904976315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,12288,0.18056533071729872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,128,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,32,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,2048,0.008014221986134848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,768,0.0042239998777707415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,4096,0.03163999981350369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,128,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,5120,0.03714311122894287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,7168,0.03322133421897888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,6144,0.24221155378553602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,256,0.046762665112813316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,2048,0.008634666601816813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,8192,0.016502221425374348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,2048,0.013350221845838757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,512,0.014442667365074158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,4096,0.014919999572965833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,768,0.0028702221396896574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,5120,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,5120,0.009296000003814697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,1536,0.004530666602982415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,1536,0.011172444456153445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,1536,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,10240,0.029678222205903795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,3072,0.015597333510716757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,6144,0.013767110804716745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,8192,0.015463999576038785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,256,0.0025315555847353404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,1536,0.008619555168681675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,65536,0.32400621308220756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,1536,0.004498666773239772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,16384,0.05611822340223524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,12288,0.03652088840802511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,10240,0.01698933376206292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,1024,0.04307111104329427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,768,0.008758222063382467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,1024,0.005962666538026597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,128,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,2048,0.006783111227883234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,4096,0.5751928753323025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,65536,0.07711911201477051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,1536,0.011034666664070554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,768,0.04791466726197136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,1536,0.023980443676312763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,512,0.015139554937680563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,1024,0.012096000214417776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,1536,0.023394667439990576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,3072,0.05492622322506375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,2560,0.02217688825395372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,1536,0.032912890116373696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,6144,0.023687111006842718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,65536,0.012396444049146442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,512,0.002887111157178879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,2048,0.03112444281578064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,65536,0.15614488389756945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,768,0.01695644524362352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,65536,0.015066666735543145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,3584,0.01567911108334859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,64,0.0038319999972979226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,512,0.01443199978934394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,7168,0.018928888771269057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,128,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,12288,0.03312177790535821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,1536,0.006384000182151794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,2560,0.006928000185224745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,10240,0.021682666407691106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,2048,0.004543999830881755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,4096,0.026966222458415564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,1024,0.014473777678277759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,768,0.005937777873542573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,2560,0.009116444322797988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,32,0.01034400032626258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,2560,0.017102221647898357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,12288,0.6724710994296603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,7168,0.029714667134814795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,128,0.015136889285511441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,16384,0.0679688917265998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,1024,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,1536,0.005929777605666055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,256,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,12288,0.01749955614407857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,8192,0.04664444592263964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,65536,0.07599733273188274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,512,0.012401777837011548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,512,0.005663110978073544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,1536,0.006197333335876465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,7168,0.02985066506597731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,16384,0.19229511419932047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,512,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,7168,0.3589182164933946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,64,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,12288,0.03461777832772996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,256,0.005561777700980504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,5120,0.043603556023703686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,5120,0.015503999259736804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,5120,0.021903110875023737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,16384,0.056233776940239795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,256,0.015472888946533203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,10240,0.015621332658661736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,16384,0.030551999807357788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,3584,0.03721333212322659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,64,0.006000888844331105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,1536,0.010737777584128909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,5120,0.010007999837398529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,16384,0.04176000091764662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,4096,0.009735110733244155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,16384,0.0741199983490838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,128,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,1024,0.012429333395428128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,256,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,3072,0.012972444295883179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,768,0.01496444476975335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,16384,0.09407822291056316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,5120,0.01846844454606374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,65536,0.07561600208282471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,10240,0.0992586678928799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,768,0.008358221915033128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,64,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,256,0.015828443898095023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,3584,0.061101330651177295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,65536,0.009335999687512716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,5120,0.11468977398342556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,1536,0.006565333240562015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,128,0.014155555102560254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,5120,0.016341333587964375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,10240,0.026944888962639704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,65536,0.10034666458765666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,512,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,64,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,256,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,2560,0.039120889372295804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,8192,0.013039111263222165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,5120,0.02971200148264567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,1536,0.04085333479775323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,3072,0.027811557054519653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,6144,0.01205244412024816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,5120,0.01517511076397366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,65536,0.3232506646050347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,512,0.006125333408514659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,8192,0.08098222149742974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,3584,0.011896000140243106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,2048,0.006589333216349284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,128,0.007636444436179266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,256,0.028540445698632136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,256,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,768,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,6144,0.06757244136598375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,1024,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,1536,0.02460622290770213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,32,0.00221422231859631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,1024,0.0034391110142072043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,8192,0.05267733335494995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,6144,0.008973333570692275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,3072,0.009644444617960189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,8192,0.011722666521867117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,2048,0.009149333669079674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,768,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,2048,0.006673777682913675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,64,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,512,0.014469333820872836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,5120,0.009122666385438707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,12288,0.018541332748201158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,12288,0.023004444109068975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,512,0.022981334063741896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,5120,0.013539555172125498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,768,0.0032435556252797446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,10240,0.016226665841208566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,512,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,7168,0.01311377767059538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,512,0.0025351111673646504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,5120,0.04049777653482225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,12288,0.04719466633266873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,256,0.007257777783605788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,8192,0.01386222243309021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,3072,0.007236444287829929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,2560,0.23226221402486166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,64,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,10240,0.2411875459882948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,512,0.0025351111673646504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,4096,0.023366222778956097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,65536,0.07825511031680636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,5120,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,2048,0.040844443771574236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,256,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,6144,0.12975377506679958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,8192,0.010695999695195092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,3584,0.027456000447273254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,65536,0.1747466723124186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,12288,0.014084445105658637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,6144,0.027471999327341717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,65536,0.017809778451919556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,512,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,12288,0.01253600004646513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,3584,0.011018666956159802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,10240,0.04754399922158983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,768,0.004576888763242298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,512,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,2560,0.014492443866199918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,2048,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,64,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,12288,0.08511377705468072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,10240,0.0177848885456721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,2048,0.006025777922736273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,3072,0.007143110864692264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,10240,0.020637333393096924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,6144,0.04079200161827935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,512,0.005888888819350137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,5120,0.004576888763242298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,3072,0.044891556104024254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,1024,0.007428444094128079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,5120,0.02068622244728936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,10240,0.017204445269372728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,1536,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,8192,0.011680888632933298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,2048,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,6144,0.01890222231547038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,3072,0.015462223026487561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,65536,0.035713778601752386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,4096,0.09402310848236084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,3072,0.007306666837798224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,5120,0.02233155568440755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,32,0.014775110615624322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,7168,0.015112888481881885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,512,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,128,0.020971554848882887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,128,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,6144,0.008976000050703684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,7168,0.023132445083724126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,7168,0.01632000009218852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,6144,0.010999110837777456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,1536,0.006621333460013072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,6144,0.01366488883892695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,2560,0.010031110710567897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,4096,0.011746666497654386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,3584,0.013387555877367655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,768,0.016812443733215332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,1024,0.015811555915408664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,16384,0.02943199872970581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,4096,0.02200800014866723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,64,0.005569777968857024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,8192,0.02130844526820713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,8192,0.02528000043498145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,4096,0.027452444036801655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,6144,0.011615110768212212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,32,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,3584,0.050272888607449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,4096,0.028655999236636694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,768,0.005950222412745158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,2560,0.011177777416176267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,8192,0.24144355456034342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,6144,0.1141244437959459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,10240,0.013818666338920593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,6144,0.008665777742862701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,2560,0.05013599991798401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,8192,0.032264000839657254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,3072,0.021401777863502502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,32,0.002485333424475458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,128,0.01313066648112403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,16384,0.021941334009170532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,64,0.002817777709828483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,7168,0.030713776747385662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,7168,0.027319999204741582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,5120,0.015631111131774057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,2560,0.00905955582857132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,512,0.00828888929552502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,65536,0.3675253391265869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,12288,0.22815201017591688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,65536,0.08630933364232381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,10240,0.07407733466890123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,6144,0.019088000059127808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,768,0.015817777978049386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,3072,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,1536,0.05130577749676175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,4096,0.0317191117339664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,3072,0.03664888938268026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,3584,0.0064693333374129395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,3584,0.007647111184067196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,7168,0.02370400064521366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,4096,0.008924444516499838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,10240,0.01201866649919086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,8192,0.017232000827789307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,5120,0.015818667080667283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,512,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,4096,0.006255110932721033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,3072,0.02923911147647434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,3584,0.015849777393870883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,1536,0.029840889904234145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,2048,0.015789333317014907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,1024,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,256,0.010650667051474253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,65536,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,1024,0.011399110986126794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,6144,0.17134400208791098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,512,0.01854222185081906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,16384,0.034167998366885714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,3072,0.010709332923094431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,768,0.01612355477280087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,2048,0.03390311201413473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,768,0.00794222205877304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,2048,0.024114666713608637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,5120,0.012421333127551608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,128,0.002874666617976295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,16384,0.03253777821858724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,4096,0.05228444602754381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,2048,0.00833511104186376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,10240,0.1914231114917331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,10240,0.023307555251651343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,768,0.0039022221333450745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,65536,0.05505777729882134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,2048,0.011748444702890186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,6144,0.012431999875439538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,5120,0.10816622442669338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,7168,0.014583999911944071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,6144,0.00904888908068339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,2560,0.007251555720965068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,2560,0.02163999941613939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,3584,0.012072000238630505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,6144,0.0042568888101312844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,64,0.0018515555808941524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,3584,0.015800888339678448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,8192,0.024863110648261175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,256,0.007259555160999298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,3584,0.007628444168302748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,16384,0.030132446024152968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,65536,0.0840302242173089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,32,0.005790222022268508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,768,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,2560,0.011200000014570026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,10240,0.03147466646300422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,512,0.012085333466529846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,768,0.01555377741654714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,2560,0.007241778075695038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,1024,0.021613332960340712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,64,0.0021760000122918025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,7168,0.2476551002926297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,768,0.0028168888141711554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,768,0.003177777760558658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,65536,0.6261742379930284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,10240,0.026912000444200303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,1536,0.013797333670987023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,6144,0.00905511114332411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,128,0.014107555150985718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,128,0.014456000592973499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,10240,0.004589333302444882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,12288,0.05336088935534159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,768,0.31075112024943036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,12288,0.011699555648697747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,7168,0.5816711319817437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,2048,0.020272889071040683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,7168,0.0411119990878635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,6144,0.26341154840257436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,1536,0.009011555876996782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,5120,0.00867555538813273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,128,0.01683644453684489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,65536,0.1426773336198595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,12288,0.028102222416136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,16384,0.02404888967672984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,16384,0.025619554850790236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,12288,0.016007110476493835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,1536,0.00601333338353369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,768,0.002850666642189026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,2048,0.12620355023278132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,7168,0.058423108524746366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,2560,0.004548444516128964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,2560,0.006650666809744305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,12288,0.02520977788501316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,7168,0.022862222459581163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,256,0.03756444321738349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,12288,0.02197333342499203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,8192,0.014411555396185981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,128,0.013453333742088742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,1536,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,65536,0.28106845749749076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,16384,0.01684977776474423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,32,0.007613333563009898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,512,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,7168,0.038576887713538274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,5120,0.03397599856058756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,768,0.007481777833567725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,7168,0.06786311335033841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,8192,0.01851111153761546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,3584,0.008299555215570662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,8192,0.33835911750793457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,128,0.007592889169851939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,64,0.005635555419656966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,10240,0.017834666702482436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,512,0.0069439998931354946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,12288,0.03813599877887302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,2560,0.007072000039948358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,65536,0.34810400009155273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,10240,0.0995031131638421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,6144,0.03154311246342129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,2560,0.004948444250557157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,10240,0.16378666294945612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,4096,0.06279377804862128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,1024,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,1024,0.01752622259987725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,2560,0.03215022219551934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,64,0.006245333287451003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,2560,0.015591111448076038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,1536,0.0069164443347189165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,2048,0.007479999628331926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,6144,0.03812088900142246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,32,0.007673777639865875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,32,0.001818666648533609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,4096,0.026871111657884385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,2048,0.07135377989874946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,512,0.002499555548032125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,64,0.0032151111712058387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,7168,0.010334221853150262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,6144,0.008345777789751688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,7168,0.009443555441167619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,12288,0.1603484418657091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,65536,0.0873475538359748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,10240,0.07164622015423246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,2048,0.006583111153708563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,32,0.0021893332401911416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,4096,0.008622222476535374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,12288,0.020632889535692003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,16384,0.019089778264363606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,512,0.00554666668176651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,32,0.007307555940416124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,512,0.012720000412729053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,10240,0.011798222031858234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,32,0.001842666624320878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,512,0.012504888905419244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,768,0.014860444598727755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,7168,0.08932533529069688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,1536,0.006335999816656113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,768,0.01447111037042406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,768,0.005014222115278244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,512,0.002536000063021978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,3072,0.021965333157115515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,6144,0.026093333959579468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,65536,0.04766044351789686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,8192,0.03283199999067519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,5120,0.0041911109454101985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,8192,0.009313777916961247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,16384,0.01514311134815216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,12288,0.18535821967654756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,65536,0.025366221865018208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,65536,0.12880622016059026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,2560,0.02712000078625149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,2048,0.0069324444565508105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,256,0.03020355436537001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,768,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,12288,0.036347554789649114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,16384,0.026738666825824316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,128,0.014094221923086377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,10240,0.03628355595800612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,12288,0.020346666375796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,32,0.004913777940803104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,7168,0.01928711103068458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,4096,0.017198221551047433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,3584,0.007330666813585494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,768,0.004935111022657818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,1536,0.013572444518407186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,3072,0.020590222544140287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,4096,0.06422489219241671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,4096,0.02760355505678389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,2560,0.007992000215583378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,128,0.040595554643207125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,1536,0.003922666526503033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,16384,0.047094222572114736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,7168,0.39251110288831925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,768,0.005373333477311664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,2560,0.015547555353906421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,7168,0.019985778464211356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,2048,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,7168,0.009768889182143742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,3584,0.012538666526476542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,65536,0.08315022124184503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,1536,0.02568977740075853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,16384,0.04623022344377306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,1024,0.0027840000887711844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,3584,0.013474666410022311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,7168,0.07791466845406426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,3072,0.014490667316648694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,2048,0.011863110793961419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,512,0.002502222235004107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,2048,0.010840000377760993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,5120,0.0181342214345932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,6144,0.03253600001335144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,3072,0.006986666884687211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,512,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,12288,0.29736532105339897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,768,0.002509333400262727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,512,0.003583111282851961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,128,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,128,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,3584,0.1060897774166531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,2048,0.008630221916569604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,3584,0.029656890365812514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,1024,0.01887022289964888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,7168,0.051460444927215576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,512,0.0024951110697454875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,2048,0.004564444637960858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,65536,0.1690737803777059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,32,0.005196444276306364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,65536,0.1169475581910875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,2048,0.07422666417227851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,256,0.02456888887617323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,256,0.00831822223133511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,5120,0.03102311160829332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,65536,0.14759555127885607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,768,0.0058648888435628675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,2560,0.012609778179062737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,5120,0.026847110854254827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,8192,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,12288,0.013733333183659447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,16384,0.02938577863905165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,1536,0.007247111035717859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,2048,0.004927999857399198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,4096,0.00795911086930169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,7168,0.019302222463819716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,256,0.002546666603949335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,3072,0.0074577778577804565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,1024,0.003549333247873518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,8192,0.05554755528767904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,7168,0.010446222292052375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,256,0.021342222889264423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,256,0.004974222017659081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,10240,0.024684444069862366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,128,0.0025804444319672054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,12288,0.5804195404052734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,256,0.01482399966981676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,12288,0.009374221993817223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,6144,0.03532088796297709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,10240,0.005993777679072485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,3584,0.03426577647527059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,256,0.0026213334252436957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,1024,0.0047120000753137796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,768,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,3584,0.011360000405046674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,8192,0.015118221441904703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,128,0.002263111165828175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,4096,0.007670222057236566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,8192,0.009396444592210982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,4096,0.012171555724408893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,1536,0.03968888852331374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,16384,0.011052444577217102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,5120,0.022671111755900916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,1536,0.013053333593739403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,16384,0.022229333718617756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,32,0.0053280000057485365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,16384,0.051819556289248996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,256,0.0053075556125905775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,1024,0.005793777604897817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,2048,0.006969778074158563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,768,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,1024,0.01850844422976176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,10240,0.01625422139962514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,768,0.003136000078585413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,3072,0.012905778156386482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,64,0.0053119998839166425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,6144,0.024714666936132643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,3584,0.00925866680012809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,1536,0.008330666356616551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,12288,0.02405066622628106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,7168,0.05141689048873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,3072,0.031629333893458046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,2048,0.027804444233576458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,2560,0.006562666760550604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,2048,0.0064151109092765385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,256,0.02036888897418976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,8192,0.037251555257373385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,1024,0.0058106668293476105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,512,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,10240,0.02445777753988902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,768,0.0028408887899584244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,6144,0.1991120047039456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,2048,0.0042160000238153665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,2048,0.07527733511394925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,768,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,8192,0.011368888947698804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,65536,10.781826443142362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,7168,0.154731551806132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,65536,0.2701031102074517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,7168,0.020232000284724765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,2048,0.008268444074524773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,512,0.003142222141226133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,64,0.0025315555847353404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,128,0.021478222476111516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,5120,0.059747556845347084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,1536,0.01651822196112739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,2048,0.014474666780895658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,1536,0.00721599989467197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,65536,0.07301777601242065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,32,0.009183111290136972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,1024,0.02624622152911292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,10240,0.02330577704641554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,10240,0.025410667061805725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,3584,0.007252444823582967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,32,0.006598222172922558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,65536,0.09444177813000149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,2560,0.009648000200589498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,2048,0.01589777734544542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,7168,1.0932142469618056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,10240,0.055530667304992676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,6144,0.025082666012975905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,256,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,32,0.0018639999131361644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,8192,0.025557332568698462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,10240,0.25824978616502553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,768,0.016191111670600045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,2048,0.006611555400821898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,10240,0.02356622285313076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,1024,0.01107022249036365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,512,0.01755644381046295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,16384,0.024642666180928547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,2560,0.021292444732454088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,16384,0.09538488917880589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,65536,0.07325066460503472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,1024,0.035897778140174016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,32,0.001983111103375753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,3584,0.010336000058386061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,32,0.0018702221827374564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,4096,0.008963555925422246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,8192,0.004247111164861255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,3584,0.006220444622966979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,3584,0.007333333293596904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,1024,0.015084443820847405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,65536,0.081258667839898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,4096,0.014682667122946845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,10240,0.05338755581114027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,5120,0.16034932931264242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,8192,0.01076533314254549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,128,0.011376000112957425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,768,0.005702222386995952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,12288,0.022824888428052265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,2560,0.006613333192136552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,768,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,7168,0.009768889182143742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,16384,0.027707555227809485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,6144,0.07493778069814046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,10240,0.034873776965671115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,12288,0.03646399908595615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,128,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,256,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,256,0.004879999905824661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,8192,0.22918489244249132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,64,0.005952000204059813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,5120,0.02986577815479702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,7168,0.043997334109412305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,8192,0.010666666759385003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,12288,0.012439111040698158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,5120,0.00833511104186376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,4096,0.007256000406212277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,1536,0.08343911170959473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,128,0.04155822263823615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,2048,0.01925066610177358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,512,0.018235555953449674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,7168,0.010690666735172272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,10240,0.0351511107550727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,64,0.005577777822812398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,512,0.0026453334010309647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,6144,0.016730666160583496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,1024,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,1536,0.017610667480362784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,128,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,512,0.005210666606823604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,2048,0.02742133372359806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,8192,0.02894044584698147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,16384,0.04656889041264852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,1536,0.027796443965699937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,3072,0.00759733302725686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,3584,0.00868622213602066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,8192,0.06076800160937839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,16384,0.012105777859687805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,7168,0.013077333569526672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,768,0.04833866821395027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,32,0.006583999842405319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,256,0.0018986666368113623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,512,0.006218666831652324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,256,0.016840888394249808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,768,0.019652444455358718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,6144,0.04993777804904514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,5120,0.024117334021462336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,65536,0.09050044748518203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,6144,0.008994667066468133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,512,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,3072,0.0144533332851198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,1536,0.0038613333470291565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,5120,0.008567111359702216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,12288,0.09746222363577949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,256,0.00519466648499171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,512,0.004658666749795278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,768,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,7168,1.1303955713907878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,3584,0.03272622161441379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,512,0.00319377767542998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,64,0.0021893332401911416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,4096,0.01312888910373052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,10240,0.01717688971095615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,768,0.004096888833575779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,12288,0.07624799675411649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,128,0.0021182222084866627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,1536,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,6144,0.014142221874660917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,1024,0.005877333382765452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,2048,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,2560,0.1011600030793084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,5120,0.008721777962313758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,16384,0.015471999843915304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,16384,0.06528800063663058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,512,0.010024000373151567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,128,0.005983110931184556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,12288,0.01769244505299462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,6144,0.01797066628932953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,6144,0.03919733232922024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,3584,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,65536,0.10704977644814385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,64,0.0018435555199782054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,5120,0.009304888546466827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,7168,0.016089777151743572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,768,0.007284444239404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,2048,0.03509066502253214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,65536,0.09114578035142686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,7168,0.017883555756674874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,2048,0.004831999954250124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,6144,0.008967999782827165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,5120,0.01036266651418474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,10240,0.020649777518378366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,1536,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,12288,0.02977777851952447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,3584,0.00831733312871721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,2048,0.013861333330472311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,512,0.010882666541470421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,256,0.005694222119119432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,4096,0.007703111403518253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,12288,0.08185866806242201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,65536,0.025282666087150574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,6144,0.03864177730348375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,768,0.005454222361246745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,3584,1.0057182312011719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,8192,0.022971555590629578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,512,0.00249422217408816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,3072,0.016751110553741455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,1024,0.007271111011505127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,4096,0.009759999811649323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,2048,0.011705777711338468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,256,0.011727111207114326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,1024,0.016538666354285348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,12288,0.6758284568786621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,5120,0.01719377769364251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,1536,0.05154133174154493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,1024,0.01348622226052814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,3072,0.007303111255168915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,128,0.004918222212129169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,32,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,3072,0.007567110988828871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,65536,0.06628088818656074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,2560,0.007279111279381647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,6144,0.015541333291265698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,1536,0.006284444282452266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,7168,0.008007999923494127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,1024,0.01718755563100179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,7168,0.013258666627936892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,64,0.004204444587230682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,5120,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,10240,0.023023110296991136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,2560,0.012452444268597497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,6144,0.019928889142142404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,1024,0.005963555640644497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,8192,0.16528088516659206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,65536,0.046000000503328115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,1536,0.011704000333944956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,64,0.006236444330877728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,3072,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,256,0.004534222185611725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,1024,0.018014222383499146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,10240,0.018597332967652213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,256,0.005881777654091517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,768,0.008055999875068665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,768,0.004137777619891697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,512,0.01551822159025404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,5120,0.008624888956546783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,7168,0.013447110851605734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,128,0.014771555860837301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,3584,0.008728000024954477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,64,0.0018648888087934917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,1024,0.012650666965378655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,4096,0.022720888257026672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,64,0.001985777790347735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,7168,0.03309155503908793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,768,0.005927111125654644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,12288,0.1255306667751736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,4096,0.027113777067926195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,512,0.002830222249031067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,4096,0.011063999599880643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,512,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,64,0.005544888890451855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,768,0.01752622259987725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,1024,0.06945955753326416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,3584,0.007336888876226213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,512,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,32,0.008976000050703684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,3584,0.025335111551814612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,2048,0.14894399378034803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,4096,0.012510221865442065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,512,0.021713778376579285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,5120,0.008991111483838823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,7168,0.01407999959256914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,7168,0.013947554760509066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,3584,0.007905777957704332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,10240,0.03053599927160475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,16384,0.015490666031837463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,10240,0.20342932807074654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,12288,0.013779555757840475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,512,0.012735111018021902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,64,0.009668444593747457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,768,0.018761777215533786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,65536,5.835692511664496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,5120,0.00903555585278405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,8192,0.38880621062384707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,512,0.007279111279381647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,2048,0.00795733349190818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,512,0.005680000202523337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,16384,0.032840887705485024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,2048,0.01295199990272522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,10240,0.02914666798379686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,512,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,6144,0.019132445255915325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,8192,0.036912888288497925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,256,0.0024826667375034755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,32,0.003181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,5120,0.004840888910823398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,12288,0.01274222218328052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,32,0.0058755555914507965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,4096,0.03052977720896403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,512,0.006261333409282897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,1024,0.014097778333557976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,32,0.0021280000607172647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,10240,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,16384,0.21886844105190703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,4096,0.007891555627187094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,5120,0.01756000022093455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,768,0.007629333270920648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,65536,0.08259911007351346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,6144,0.01755555636352963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,12288,0.012655999925401477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,5120,0.03353510962592231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,10240,0.01761155492729611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,2560,0.007220444579919179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,1024,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,256,0.004869333157936732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,12288,0.018187556001875136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,3072,0.031424889961878456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,3584,0.011732444167137146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,3584,0.02852533260981242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,2048,0.012786666552225748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,128,0.005509333478079901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,7168,0.010253333383136326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,768,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,2560,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,512,0.0028577778074476454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,1024,0.003132444495956103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,7168,0.011485333244005838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,256,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,3072,0.11388888623979355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,7168,0.0144177774588267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,4096,0.04408533374468485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,7168,0.014840000205569796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,7168,0.019866666860050626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,65536,0.25721245341830784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,12288,0.030320889419979517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,8192,0.027816888358857896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,768,0.009312888814343346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,3072,0.007272889216740926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,3584,0.016158221496476066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,3072,0.014115555418862237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,10240,0.023541332946883306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,8192,0.025701334079106648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,1024,0.00573511090543535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,16384,0.023031999667485554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,512,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,2048,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,5120,0.01204355557759603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,3072,0.010435555544164447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,8192,0.028252445989184912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,768,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,2560,0.02411288850837284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,8192,0.14344622029198542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,1536,0.00629688882165485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,2560,0.07655644416809082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,5120,0.1475342247221205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,4096,0.009929777847396003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,768,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,256,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,256,0.02237155536810557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,512,0.006215111249023014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,256,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,2048,0.01556711064444648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,4096,0.2219555510414971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,1024,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,10240,0.01204533295498954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,8192,0.02398577829202016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,4096,0.007768889268239339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,2560,0.008435555630260045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,4096,0.03609866566128201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,4096,0.007930667036109501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,7168,0.01311644415060679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,2048,0.006254222244024277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,10240,0.05144355694452921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,3072,0.00868800034125646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,3584,0.006452444526884291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,10240,0.04784444305631849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,1536,0.02260977857642704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,65536,0.06432444519466825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,1024,0.009276444713274637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,6144,0.010847110715177325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,512,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,6144,0.016512889001104567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,4096,0.030993776188956365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,2048,0.00796888851457172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,256,0.004920888692140579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,12288,0.2305875619252523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,1536,0.005924444645643234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,2048,0.01274222218328052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,6144,0.013777777552604675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,768,0.015450666348139444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,2560,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,512,0.016559110747443307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,7168,0.02849422229660882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,2560,0.10476977957619561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,65536,0.41176356209648984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,2048,0.006250666661394968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,10240,0.018834667073355783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,8192,0.03568977779812283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,12288,0.013806222213639153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,64,0.006608888920810487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,512,0.01275199982855055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,1024,0.007256888680987888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,3072,0.007290666302045186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,2560,0.010392000277837118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,256,0.002495999965402815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,3072,0.015461333923869662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,16384,0.015909334023793537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,8192,0.04607377780808342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,4096,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,6144,0.070342222849528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,1024,0.018583999739752877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,1536,0.006904888898134232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,1536,0.003928889003064898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,8192,0.020627554919984605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,3072,0.014836443795098199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,7168,0.010424888796276517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,3072,0.007092444433106317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,3072,0.009700444837411245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,10240,0.05184622274504768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,7168,0.03303644392225478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,2560,0.12675644291771784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,12288,0.04895111256175571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,10240,0.012039110892348819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,7168,0.01962488889694214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,768,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,5120,0.022623111804326374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,16384,0.06055821975072225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,3584,0.04051911168628269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,2560,0.006879999819729064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,768,0.008550222549173567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,256,0.002566222308410539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,128,0.022639110684394836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,3584,0.008019555773999956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,1536,0.013832888669437833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,10240,0.025457777910762366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,1536,0.006250666661394968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,12288,0.023932443724738225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,16384,0.024404444628291663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,3072,0.0440577765305837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,768,0.00591822216908137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,3072,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,32,0.001790222194459703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,65536,0.015448000695970325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,768,0.018224888377719455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,5120,0.015469332536061605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,3072,0.012543111211723752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,1024,0.029049777322345312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,768,0.007988444632954067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,7168,0.009992889232105678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,5120,0.02134311033619775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,512,0.006235555642180973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,5120,0.010413332945770687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,7168,0.05936444467968411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,128,0.004588444613748127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,1024,0.06070578098297119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,32,0.004544888933499654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,256,0.006903111106819577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,1536,0.07272089189953275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,7168,0.046132445335388184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,64,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,2048,0.022656889425383672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,3072,0.007246221933099959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,65536,0.07501333289676242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,768,0.002842666581273079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,32,0.003157333367400699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,10240,0.01848088867134518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,1536,0.0068604445291890045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,3072,0.016546666622161865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,3072,0.018533334136009216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,10240,0.008999110923873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,128,0.010024000373151567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,12288,0.034437331888410784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,12288,0.2372844484117296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,6144,0.02892177634769016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,3584,0.04260177744759453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,768,0.00590133335855272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,5120,0.024088000257809956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,10240,0.006666666517655055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,10240,0.013391999734772576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,8192,0.029158221350775823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,128,0.00217333332531982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,128,0.014783110883500842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,256,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,1024,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,4096,0.006302222195598815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,16384,0.027202667461501226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,768,0.006806222101052602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,512,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,2048,0.04134044382307265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,12288,0.021172444025675457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,512,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,2560,0.0876728892326355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,8192,0.010058666268984476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,2048,0.022852443986468848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,12288,0.5506586498684353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,512,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,1536,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,4096,0.09708888663185967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,65536,0.16683288415273032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,1024,0.00795911086930169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,1536,0.01716800034046173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,5120,0.008314666648705801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,128,0.013409777647919126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,5120,0.009712888962692684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,256,0.0023208889696333143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,7168,0.011699555648697747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,12288,0.01958755486541324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,32,0.0021128888345426982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,8192,0.020249777370029025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,1536,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,2048,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,4096,0.09743289152781169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,2048,0.006226666685607698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,768,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,3072,0.012331555287043253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,10240,0.04215822286076016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,64,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,768,0.00281866660548581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,3584,0.0063182223174307085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,64,0.005719999886221356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,768,0.005667555663320754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,64,0.005939555664857228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,7168,0.02880000074704488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,5120,0.008618666893906063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,10240,0.13110400570763484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,6144,0.017100445098347135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,128,0.002986666642957263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,5120,0.03644711110326979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,5120,0.009343110852771336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,65536,0.0938639971945021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,3072,0.025807110799683466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,12288,0.011792889071835412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,2560,0.013380444712109037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,2048,0.013080000049538083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,128,0.004875555634498596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,512,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,256,0.014095111025704278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,16384,0.029089778661727905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,3584,0.010067555639478896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,256,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,3584,0.011351111034552256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,7168,0.021239999267790053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,512,0.011398221883508893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,6144,0.012160888976520963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,1024,0.01995377739270528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,12288,0.004675555560323927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,5120,0.016493333710564505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,1024,0.00590844452381134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,6144,0.004886222382386525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,3584,0.01426577733622657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,3072,0.05737777551015218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,128,0.012479110724396177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,3072,0.007283555964628856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,8192,0.007661333514584436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,10240,0.03427022364404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,65536,0.22585156228807238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,3072,0.014651555154058667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,2560,0.0069857777820693115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,4096,0.015087111128701104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,256,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,2560,0.02942488922013177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,3072,0.007275555696752336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,512,0.006584888945023219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,1024,0.007244444555706448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,256,0.004889777965015835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,256,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,1024,0.025815111067559984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,5120,0.023623110519515142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,8192,0.015810666812790763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,5120,0.017623111605644226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,512,0.0058755555914507965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,2560,0.011037333144081963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,128,0.0019155556543005838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,16384,0.03487288951873779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,4096,0.00793866647614373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,256,0.002485333424475458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,3584,0.01034577770365609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,3072,0.4442417886522081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,768,0.006945778098371293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,7168,0.02805333336194356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,16384,0.03505333264668783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,65536,0.07467289103402032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,65536,0.10272355874379475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,128,0.002159111201763153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,128,0.023994666006830003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,6144,0.05120977759361267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,5120,0.032920890384250216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,4096,0.022446221775478784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,16384,0.01707466608948178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,65536,0.07321244478225708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,6144,0.029608001311620075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,10240,0.019191111127535503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,1536,0.012818666795889536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,10240,0.02027466727627648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,768,0.0069439998931354946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,128,0.006981333096822103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,4096,0.35094488991631406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,3584,0.005632888939645555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,7168,0.0440577765305837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,2048,0.01443733274936676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,32,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,8192,0.03438933359252082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,512,0.004237333519591225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,768,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,10240,0.10040799776713054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,1024,0.007975999679830339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,6144,0.009005332986513773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,6144,0.016062221593326993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,128,0.005931555396980709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,1024,0.011674666570292579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,1024,0.015397333436542086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,5120,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,1024,0.0031742221779293488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,1536,0.005296000176005893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,128,0.005607999861240387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,2560,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,12288,0.05794577466117012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,5120,0.027468444572554693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,8192,0.03692533241377936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,128,0.009354666703277165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,12288,0.021311110920376245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,3584,0.015476443701320224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,6144,0.15390843815273708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,10240,0.01298577752378252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,256,0.006251555350091722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,2048,0.024506666594081458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,6144,0.03084889054298401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,256,0.013304000099500021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,7168,0.12783200210995144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,128,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,8192,0.010782221953074137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,256,0.0021128888345426982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,3072,0.015488889482286243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,16384,0.01683911184469859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,2048,0.03139466709560818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,8192,0.014950222439236112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,6144,0.007859555383523306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,2560,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,512,0.010071111222108206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,1024,0.0028391112056043413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,1024,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,4096,0.041082666979895696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,64,0.0021155555215146807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,1536,0.028770665327707928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,65536,0.08565866947174072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,16384,0.016154666741689045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,10240,0.03503555390569899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,65536,0.04229066769282023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,16384,0.2603813277350532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,3584,0.018527110417683918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,256,0.016504888733228047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,3584,0.1892391045888265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,2560,0.0064071110553211635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,4096,0.00829777783817715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,1536,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,64,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,2560,0.006631111105283101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,512,0.02032533288002014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,256,0.00834222220712238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,3584,0.01424888935354021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,65536,0.08465511269039577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,10240,0.018515555395020377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,768,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,6144,0.013285333083735572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,2560,0.013973333769374423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,8192,0.01406311160988278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,3072,0.005535999933878581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,1536,0.010577778021494547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,2048,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,5120,0.04641511042912801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,6144,0.13611111376020643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,1024,0.007291555404663086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,128,0.015558222929636637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,65536,0.07098044289482965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,64,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,3072,0.00701422244310379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,6144,0.015133332875039844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,6144,0.16125510798560247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,4096,0.1029706663555569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,1536,0.013368888861603208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,2048,0.006552000012662675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,3072,0.028168890211317275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,1536,0.006022222340106964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,8192,0.05673155519697401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,128,0.008626666333940294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,10240,0.03630933496687148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,256,0.0025333333760499954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,1024,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,64,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,1536,0.008357333640257517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,5120,0.004920000003443824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,16384,0.05255022313859728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,12288,0.02682222260369195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,3072,0.007024889190991719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,16384,0.02529955572552151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,256,0.006515555497672822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,1536,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,6144,0.058555556668175586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,256,0.0018586667461527719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,256,0.007629333270920648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,10240,1.962088902791341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,3584,0.0196088883611891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,512,0.005422222117582957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,512,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,4096,0.012778666284349231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,2560,0.007328888608349695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,2560,0.006981333096822103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,768,0.0032355555643637977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,768,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,2048,0.028282665544086035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,2048,0.006305777778228124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,128,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,128,0.00453955555955569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,256,0.030429333448410034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,65536,0.08349066972732544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,512,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,2560,0.011367110742463006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,10240,0.03391999999682108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,1024,0.01107022249036365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,4096,0.01368977791733212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,2560,0.030673778719372217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,4096,0.01890222231547038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,10240,0.016875555117925007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,10240,0.09954222043355306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,4096,0.00793688909875022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,2560,0.041944887903001576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,4096,0.012753778033786349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,3072,0.03678311241997613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,768,0.00981866651111179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,512,0.005331555588377847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,3584,0.10088177522023518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,768,0.014871110518773397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,1024,0.0063235556913746735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,128,0.02568711174858941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,2560,0.007281777759393056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,6144,0.024799111816618178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,7168,0.009909333454238044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,512,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,7168,0.010424888796276517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,5120,0.012422222230169507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,8192,0.019816888703240287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,10240,0.05217688944604662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,10240,0.014072888427310519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,2048,0.02237866653336419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,5120,0.012065778175989786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,6144,0.06005777915318807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,512,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,12288,0.016415999995337594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,1536,0.00832266691658232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,1536,0.0069022224181228215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,8192,0.021682666407691106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,128,0.0069324444565508105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,64,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,1536,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,8192,0.011011555790901184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,32,0.002003555496533712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,512,0.03336711062325372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,1024,0.01332800007528729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,8192,0.010354667074150508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,512,0.016857778032620747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,3584,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,65536,0.6100311279296875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,2048,0.02299555473857456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,128,0.030212445391549006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,12288,0.032927112446890935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,1024,0.025410667061805725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,6144,0.01035733355416192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,5120,0.009093333449628618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,8192,0.014813333749771118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,65536,0.04569422205289205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,2560,0.015478221906556023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,4096,0.007836444510353936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,10240,0.01663733356528812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,3072,0.007306666837798224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,8192,0.03173066510094537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,6144,0.012065778175989786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,12288,0.01312444441848331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,8192,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,1536,0.004563555535342959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,3584,0.031304889255099826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,10240,0.033618665403789945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,5120,0.00833155545923445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,768,0.005582222094138463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,3584,0.031181332137849595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,5120,0.045223110251956515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,256,0.006843555718660355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,1024,0.01148444414138794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,4096,0.012191111014948951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,3584,0.026405332816971674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,8192,0.034648001194000244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,256,0.007662222617202335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,512,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,5120,0.009700444837411245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,128,0.015013333823945789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,768,0.015825778245925903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,65536,0.0880933337741428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,128,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,6144,0.028536889288160536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,4096,0.015560888581805758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,2560,0.006584888945023219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,12288,0.004873777843183941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,2048,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,1024,0.015186667442321777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,3584,0.00757422215408749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,4096,0.09585155381096734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,512,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,5120,0.036828445063696966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,128,0.004870222260554631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,2048,0.02278577784697215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,65536,0.06276355849372016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,8192,0.20133422480689156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,65536,0.06578133503595988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,768,0.008986666798591614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,512,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,256,0.004214222232500712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,768,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,16384,0.3428560098012288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,512,0.011378666592968835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,2560,0.01640977793269687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,3584,0.007948444121413762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,512,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,65536,0.12026933828989665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,3072,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,16384,0.15225066079033747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,65536,0.07592888673146565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,6144,0.022636445032225713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,1024,0.0041333333485656315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,2048,0.05022222134802076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,1024,0.02975377771589491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,2560,0.016880000631014507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,1536,0.00665866666369968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,3072,0.007254222200976477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,5120,0.01704355577627818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,4096,0.017861333158281114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,6144,0.02609511050913069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,32,0.005622222191757626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,5120,0.008645333349704742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,6144,0.010392888552612728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,6144,0.008902221918106079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,2560,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,4096,0.008976000050703684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,2560,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,2560,0.015747555428081088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,1536,0.008871999879678091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,3584,0.016007110476493835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,5120,0.005946666830115848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,4096,0.011414222419261932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,2048,0.016733333468437195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,3072,0.007319110963079665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,32,0.00865955568022198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,1536,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,65536,0.3457511001163059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,3584,0.00793866647614373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,3584,0.015809777710172866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,2560,0.009345778160625035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,768,0.01758577757411533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,16384,0.02019466625319587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,1536,0.007790221936172909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,10240,0.014203555054134794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,1536,0.007914666500356462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,2048,0.018976888722843595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,8192,0.01455200049612257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,256,0.012383999923865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,3072,0.07819110817379422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,1024,0.005954666684071223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,7168,0.010371555884679159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,2048,0.014929778046078153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,2560,0.006959999601046245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,8192,0.04544533292452494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,128,0.014107555150985718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,256,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,65536,0.198052446047465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,1536,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,2048,0.020985777179400127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,1024,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,8192,0.017904000149832833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,512,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,16384,0.023891554938422307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,65536,0.23221066262986925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,1536,0.09112266699473064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,256,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,128,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,6144,0.01719288859102461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,12288,0.06432444519466825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,2048,0.014497778481907315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,7168,0.0775760014851888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,1536,0.006667555620272954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,12288,0.021881777379247878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,8192,0.028926223516464233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,512,0.014864888456132678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,16384,0.02265600032276577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,4096,0.01072177787621816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,8192,0.01311288856797748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,512,0.005615111026499007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,768,0.006079999936951532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,3072,0.018562666243977018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,768,0.008651555412345463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,256,0.00981422182586458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,2048,0.01685422162214915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,5120,0.008328888979223039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,12288,0.013047110703256396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,12288,0.2670088873969184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,128,0.013079110946920184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,7168,0.038089778688218855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,4096,0.05478577812512716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,5120,0.01755822201569875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,4096,0.007968000239796108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,16384,0.06657333506478204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,3584,0.010408889088365767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,2048,0.004184000194072723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,3584,0.023406222462654114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,6144,0.00869511150651508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,3584,0.10430044598049587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,1536,0.006264888991912206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,3584,0.02606222199069129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,768,0.014782221780882942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,7168,0.018952889574898612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,64,0.0019262221952279408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,2560,0.008926221893893348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,3072,0.02568977740075853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,5120,0.03972622089915805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,16384,0.03169066707293192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,512,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,3072,0.007285333342022366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,65536,0.08351377646128337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,128,0.014120888378885059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,65536,0.08619466755125259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,6144,0.020635555187861126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,1024,0.003170666595300039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,10240,0.02130311065249973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,12288,0.010553778045707278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,6144,0.013728888498412238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,32,0.001840888833006223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,3072,0.02334755493534936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,768,0.023429332507981196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,1024,0.00565244464410676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,3072,0.023569777607917786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,8192,0.05084266596370273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,3584,0.023728888895776536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,1536,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,65536,0.04722311099370321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,2560,0.042392889658610024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,2048,0.006562666760550604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,12288,0.08859911229875352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,2048,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,512,0.006605333338181178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,512,0.00253866674999396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,16384,0.05296711126963297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,512,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,16384,0.008915555146005418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,3072,0.06740888622072008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,6144,0.013805333111021252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,3072,0.016525333126386006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,768,0.042097777128219604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,512,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,12288,0.031091554297341242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,512,0.01956977778010898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,5120,0.06663999954859416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,128,0.009693333672152625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,10240,0.022102221846580505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,3584,0.011120888921949597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,10240,0.029313776228162978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,3584,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,3584,0.00907733374171787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,256,0.014131555954615274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,5120,0.010026666853162978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,2560,0.006630222416586346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,512,0.014140443669425117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,32,0.002024000096652243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,3072,0.013083555632167392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,32,0.006994666324721442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,6144,0.004604444321658877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,10240,0.03923199905289544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,3072,0.014544000228246054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,5120,0.01147822207874722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,32,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,5120,0.04467644294102987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,768,0.00960977789428499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,32,0.012431999875439538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,64,0.002120888895458645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,4096,0.0276826669772466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,512,0.0028880000528362063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,768,0.018550222118695576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,65536,0.049386666880713574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,2048,0.006312888943486744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,3072,0.029509332444932725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,7168,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,12288,0.10496266682942708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,4096,0.011672889192899069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,128,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,5120,0.03613866700066461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,8192,0.013119999733236102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,16384,0.02347999976740943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,3072,0.0048942222363419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,5120,0.04009777638647292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,10240,0.005620444400442972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,5120,0.021931555536058214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,256,0.006302222195598815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,768,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,1024,0.02532533307870229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,1024,0.005583111196756363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,4096,0.006554666492674086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,3584,0.016760889026853774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,4096,0.023078221413824294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,1536,0.006913777854707506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,768,0.005618666609128316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,3072,0.009549332989586724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,2048,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,5120,0.008544888761308458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,32,0.00553955551650789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,512,0.013437333206335703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,1024,0.007027555671003129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,12288,0.03377688924471537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,128,0.023590222001075745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,10240,0.015390222271283468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,5120,0.008358221915033128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,512,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,512,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,3584,0.007979555262459649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,2560,0.019894222418467205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,16384,0.006589333216349284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,6144,0.012129777835475074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,7168,0.01441511180665758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,16384,0.059043553140428334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,128,0.004671999977694618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,1536,0.04044444362322489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,128,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,32,0.0025724443710512584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,65536,0.05347733365164863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,512,0.0028817777832349143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,6144,0.009787555370065901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,3072,0.040968888335757785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,3584,0.005975111077229182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,7168,0.025052444802390203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,5120,0.011762667033407422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,12288,0.018179555733998615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,4096,0.10357155402501424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,5120,0.008349333372380998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,1536,0.005872888697518243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,1024,0.005620444400442972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,1536,0.006170666466156642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,768,0.005956444475385878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,5120,0.030000001192092896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,65536,0.05773599942525228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,256,0.002201777779393726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,2560,0.022642667094866436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,6144,0.09005510807037354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,128,0.005021333280536863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,5120,0.008997333546479544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,768,0.023346667488416035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,1024,0.01684888866212633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,3072,0.039451556073294744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,7168,0.017126222451527912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,512,0.005172444300519096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,1536,0.019913777709007263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,16384,0.044377777311537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,3584,0.0064080001579390625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,3584,0.019545776976479422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,128,0.005099555684460534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,32,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,512,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,16384,0.025575111309687298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,1024,0.025048888391918604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,2048,0.006844444407357111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,3072,0.024678222007221643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,12288,0.06115021970536974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,3584,0.019917334119478863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,7168,0.02128622266981337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,6144,0.02864977717399597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,5120,0.011857777833938599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,65536,0.020250666472646926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,3584,0.011075555450386472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,3584,0.027160889572567407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,16384,0.02212711175282796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,12288,0.0276826669772466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,2560,0.022521777285469904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,256,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,3584,0.025794666674402025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,5120,0.5589280128479004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,7168,0.020015110572179157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,65536,1.972191916571723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,2560,0.0069253332912921906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,4096,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,8192,0.01865333318710327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,256,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,6144,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,256,0.002480888946188821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,3584,0.018111111389266122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,128,0.00591822216908137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,5120,0.008559111091825698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,3072,0.007271111011505127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,3584,0.00759733302725686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,32,0.002206222257680363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,4096,0.008143111235565608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,2560,0.00666933341158761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,3072,0.007256888680987888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,256,0.012293332980738746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,4096,0.009352888498041365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,4096,0.04345600141419304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,4096,0.01997511088848114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,128,0.013415111435784234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,64,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,4096,0.0075866662793689305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,12288,0.08848266469107734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,7168,0.01719377769364251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,5120,0.06760622395409478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,64,0.0035573335157500375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,256,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,5120,0.01312444441848331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,12288,0.025075554847717285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,256,0.014558222558763293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,768,0.018023111753993563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,256,0.007999111380841997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,5120,0.024088889360427856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,32,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,2560,0.040376888381110296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,1536,0.008126222425036961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,768,0.02030755579471588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,1536,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,6144,0.010710222025712332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,65536,0.055440889464484326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,16384,0.01484266585773892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,5120,0.008280889027648503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,65536,0.0888026687833998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,128,0.02457422183619605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,8192,0.023023999399609033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,12288,0.04105510976579454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,12288,0.013807999591032663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,4096,0.0079297779334916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,256,0.005134222408135732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,1024,0.11795732710096572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,65536,0.06550666358735827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,1536,0.005753777921199799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,1024,0.03130044539769491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,10240,0.030407110850016277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,768,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,3072,0.016398222910033334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,6144,0.022264889544910852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,12288,0.037011557155185275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,3072,0.00490133340160052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,1536,0.006303999986913469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,7168,0.020632000433074105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,256,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,7168,0.016750221451123554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,1536,0.0042071110672420925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,512,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,2560,0.0151555554734336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,5120,0.030768887864218816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,10240,0.016102222932709586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,65536,0.0758684476216634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,5120,0.03363733490308126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,10240,0.07010222143597074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,128,0.005635555419656966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,128,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,2048,0.009000889129108852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,5120,0.042076445288128324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,65536,0.08085422383414374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,16384,0.020267556111017864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,65536,0.07536355654398601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,256,0.005277333160241445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,10240,0.012459555433856117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,7168,2.1523119608561196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,3584,0.025719111164410908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,128,0.011047110789351992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,3584,0.012259555359681448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,128,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,6144,0.015953777564896476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,1024,0.018625777628686693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,512,0.014748444159825643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,12288,0.13011466132269964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,12288,0.02943377693494161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,16384,0.02033422225051456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,64,0.15944711367289224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,768,0.017176000608338248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,2048,0.00453333349691497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,3072,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,256,0.006574222197135289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,6144,0.012263110942310758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,12288,0.05951199928919474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,3584,0.010300444232092964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,7168,0.01926933394538032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,512,0.005249777601824866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,12288,0.012878222597969903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,64,0.0019253332995706135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,3584,0.009931555224789513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,2048,0.015798222687509324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,4096,0.03271911210483975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,3072,0.02943377693494161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,1024,0.007615110940403408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,6144,0.008975110948085785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,2560,0.010038221875826517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,12288,0.054505778683556445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,256,0.006597333484225803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,128,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,5120,0.03127200073666043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,2048,0.023491554790072974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,2560,0.02509244448608822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,1536,0.016917333006858826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,8192,0.08402844270070393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,2560,0.020336000455750358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,4096,0.01408088869518704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,16384,0.05940622091293335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,4096,0.023758222659428913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,65536,0.04006310966279771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,1024,0.021024000313546922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,3072,0.27096088727315265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,128,0.004936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,65536,0.3739688926272922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,2048,0.006264888991912206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,4096,0.02998044424586826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,10240,0.017512000269360013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,5120,0.027230223019917805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,2560,0.009027555584907532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,128,0.013716444373130798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,1024,0.0709822244114346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,128,0.011039111349317761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,6144,0.008949333594905006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,4096,0.01718222267097897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,128,0.015108444624476962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,65536,0.08128177457385592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,256,0.00517866677708096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,12288,0.014145778285132514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,2560,0.07359466950098674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,256,0.04061333338419596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,4096,0.07022933165232341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,2048,0.006628444625271692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,7168,0.010370666782061258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,3584,0.00798044436507755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,6144,0.0315048893292745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,512,0.0103511114915212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,256,0.0024782222592168385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,5120,0.009349333743254343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,8192,0.01034755590889189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,7168,0.1057422227329678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,64,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,2048,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,7168,0.014512888259357877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,4096,0.006585777633719974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,256,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,512,0.004855999930037392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,5120,0.010545777777830759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,8192,0.02985066506597731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,512,0.019221333993805777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,512,0.00655644428398874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,512,0.04717688759167989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,768,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,3584,0.007035555938879649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,3072,0.07960889074537489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,65536,0.16466132799784342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,16384,0.05377066797680325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,4096,0.00832622249921163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,3072,0.009653333160612319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,1024,0.00564533347884814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,1024,0.021223111285103693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,4096,0.016674665941132438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,512,0.029740442832310993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,8192,0.1928008927239312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,1024,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,256,0.004570666700601578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,2560,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,1024,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,1536,0.003559111307064692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,1536,0.006127999888526068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,64,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,3072,0.007266666326257918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,1024,0.0038906666967603895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,10240,0.015623999966515435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,1536,0.06645066870583428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,3584,0.010752889017264048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,128,0.005601777798599667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,1536,0.018519111805491976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,2048,0.03006311257680257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,1024,0.00554666668176651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,1536,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,768,0.0032231110251612137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,8192,0.015135111080275642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,4096,0.023751111494170293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,6144,0.1571484406789144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,2560,0.03541244400872125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,2560,0.00693866651919153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,6144,0.06382489204406738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,1024,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,12288,0.013812444276279874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,8192,0.018174222773975797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,32,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,32,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,5120,0.015061333775520325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,1024,0.003165333428316646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,768,0.01203644441233741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,512,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,8192,0.013580444786283704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,512,0.013792888985739814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,10240,0.022132444712850783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,512,0.012432888978057437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,7168,0.04212977819972568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,512,0.017656889226701524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,2560,0.011354666617181567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,2048,0.011142222417725457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,8192,0.029050668080647785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,2560,0.0073928890956772705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,3584,0.046329776446024575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,8192,0.02388800018363529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,65536,0.07389155361387464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,256,0.005208000126812193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,8192,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,16384,0.02938400043381585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,32,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,3072,0.05289599961704678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,32,0.005892444401979446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,7168,0.013042666845851474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,768,0.02918400035964118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,2048,0.03222666515244378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,65536,0.43698755900065106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,256,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,256,0.009100444614887238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,7168,0.016694222887357075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,7168,0.005521777603361342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,7168,0.009659555223253038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,3584,0.01000711073478063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,7168,0.06418666574690077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,256,0.04509422183036804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,6144,0.012461333639091916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,5120,0.010680889089902243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,7168,0.06605422496795654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,7168,0.04736622174580892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,256,0.0051893335249688895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,2560,0.03648089038001166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,7168,0.023402666052182514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,8192,0.013802666631009845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,128,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,7168,0.005917333480384614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,12288,0.06723555591371325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,16384,0.028088887532552082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,12288,0.02021511064635383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,7168,0.025990222891171772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,3072,0.006976000136799282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,10240,0.020678222179412842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,2560,0.03711466656790839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,512,0.008981333010726506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,5120,0.02095111045572493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,8192,0.015824889143308003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,256,0.012770666844315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,16384,0.363522662056817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,3584,0.01615111033121745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,512,0.004325333154863781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,3584,0.010425777898894416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,8192,0.011000889043013254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,16384,0.011598221957683563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,8192,0.02091466718249851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,768,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,32,0.005596444424655702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,12288,0.06569777594672309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,3584,0.007683555285135905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,10240,0.04310489032003614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,6144,0.010491555763615502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,8192,0.032475554280810885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,768,0.005818666683302985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,10240,0.02161600026819441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,2560,0.025556445121765137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,4096,0.004293333325121137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,4096,0.007633777956167857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,128,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,8192,0.024141333169407312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,256,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,128,0.02103288968404134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,4096,0.007143999967310164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,6144,0.0954657793045044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,8192,0.0041964443193541635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,6144,0.023592000206311543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,16384,0.00943377779589759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,6144,0.8496373494466146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,2560,0.006594666590293248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,12288,0.01993333299954732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,512,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,4096,0.5727253490024143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,32,0.0021271111650599376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,128,0.017888888716697693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,32,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,6144,0.030167112747828167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,12288,0.022309333086013794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,6144,0.020245333512624104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,8192,0.01808088852299584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,1024,0.008854221966531541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,3072,0.011337777806652917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,6144,0.008981333010726506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,1536,0.028575999869240656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,3584,0.00759733302725686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,3584,0.004239999999602635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,12288,0.0069288888739215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,16384,0.022984888818528917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,7168,0.012083555261294046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,6144,0.05609333515167236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,4096,0.004221333397759331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,8192,0.024192000428835552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,2048,0.006659555352396435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,4096,0.1210213369793362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,4096,0.007898666792445714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,8192,0.06585333082411024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,8192,0.01740977830357022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,2560,0.004882666799757215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,5120,0.0824106666776869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,6144,0.014372444815105863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,3584,0.013693333499961428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,7168,0.013423110875818463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,3584,0.02197066611713833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,768,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,12288,0.03491466575198703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,32,0.009024889104896123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,768,0.003149333306484752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,12288,0.036475555764304265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,1536,0.00658044425977601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,10240,0.10973333650165135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,6144,0.014568888478808932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,768,0.004862222406599256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,512,0.011703111231327057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,6144,0.016192888220151264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,6144,0.022611555125978258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,7168,0.013792888985739814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,128,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,6144,0.008971555365456475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,7168,0.013070222404268054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,8192,0.01609511176745097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,1536,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,2560,0.015230221880806817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,7168,0.012252444194422828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,16384,0.10206844409306844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,16384,0.0871466663148668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,2048,0.006924444602595435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,4096,0.01724799970785777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,3072,0.008548444343937768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,65536,0.08273422055774264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,5120,0.008391111261314815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,768,0.008650666309727563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,16384,0.02106222179200914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,256,0.014485332700941296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,1024,0.0058702222175068325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,4096,0.012263110942310758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,7168,0.013598221871587964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,8192,0.12193244033389622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,16384,0.07513599925571017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,8192,0.019690665933820937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,5120,0.010964444114102257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,1536,0.005975111077229182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,3072,0.07898666461308797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,10240,0.24032176865471733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,256,0.014113777213626437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,512,0.002489777695801523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,2048,0.026942221654786006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,512,0.010784000158309937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,1024,0.006281777802440856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,5120,0.00998577806684706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,128,0.005561777700980504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,65536,0.05440444416469998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,10240,0.20676622125837538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,1024,0.017864889568752713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,65536,0.11995111571417914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,16384,0.01959111127588484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,12288,0.05346044566896227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,5120,0.012440889245933957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,3584,0.007970666719807519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,4096,0.013737777868906656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,2560,0.008339555727110969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,6144,0.03648089038001166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,12288,0.02609066665172577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,1536,0.013404444687896304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,12288,0.03456177645259433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,7168,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,12288,1.384687105814616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,1024,0.006258666515350342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,16384,0.03711466656790839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,65536,0.07552178038491143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,7168,0.028203556934992473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,7168,0.03082488973935445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,8192,0.03181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,4096,0.010400888820489248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,12288,0.017070222232076857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,3584,0.010009778042634329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,65536,0.05875821908315023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,10240,0.04571377899911669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,768,0.016184889607959323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,10240,0.020961778031455144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,512,0.003118222165438864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,1024,0.007254222200976477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,7168,0.009663999908500248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,8192,0.025217778152889673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,1024,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,3072,0.007607999775144789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,1024,0.01349066694577535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,6144,0.019713777634832595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,3072,0.032277332411872015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,1024,0.012876444392734103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,5120,0.01016088906261656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,8192,0.2729964521196154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,4096,0.007677333222495184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,1024,0.006263111200597551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,6144,0.005647111270162795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,4096,0.0162124451663759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,64,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,65536,0.07250044743220012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,3584,0.00794222205877304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,3584,0.13259466489156088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,10240,0.42411287625630695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,2560,0.02536533276240031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,1024,0.005924444645643234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,512,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,128,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,12288,0.060736000537872314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,512,0.003671110918124517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,3584,0.0075777777367168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,32,0.0018542222678661346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,1024,0.016792888442675274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,512,0.01482488877243466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,2560,0.006839111033413146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,2048,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,7168,0.01685333251953125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,2560,0.013073777986897362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,1536,0.016585777203241985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,4096,0.007983999947706858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,4096,0.007996444072988298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,512,0.008048888709810045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,5120,0.01891999940077464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,6144,0.020418667131000094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,7168,0.027552889453040227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,4096,0.007998222278224098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,5120,0.01718755563100179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,3072,0.013838222457302941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,2048,0.01643288963370853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,32,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,768,0.013436444103717804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,8192,0.015419556034935845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,4096,0.010945777926180096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,256,0.0058746664888328975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,8192,0.02165066699186961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,10240,0.24505421850416395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,10240,0.011367110742463006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,7168,0.03891911109288534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,1536,0.031572444571389094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,1536,0.01648533344268799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,7168,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,1536,0.014838222000333997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,2560,0.02167911165290409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,64,0.0018568889548381169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,2560,0.4311084482404921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,1024,0.005631111148330901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,2560,0.00703022215101454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,3072,0.008661333057615492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,512,0.012488889197508493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,3072,0.0277146663930681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,4096,0.010031110710567897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,4096,0.009646221995353699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,512,0.0029537777105967202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,512,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,256,0.016195555528004963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,512,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,32,0.03977244430118137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,16384,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,10240,0.028720888826582167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,256,0.04686044322119819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,768,0.006138666636413998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,1024,0.06129333045747545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,12288,0.008273777862389883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,512,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,1024,0.005877333382765452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,3072,0.02184711065557268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,768,0.015071110592948066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,7168,0.011405333048767514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,1536,0.0035253332720862497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,12288,0.04343733191490173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,5120,0.05560533205668131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,3584,0.01368711143732071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,3072,0.023399111297395494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,1536,0.006096000058783426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,32,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,12288,0.033562666840023465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,128,0.010675555302037133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,64,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,768,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,6144,0.009462222456932068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,7168,0.015799111790127225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,768,0.013157332936922709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,1536,0.005995555470387141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,4096,0.03770844472779168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,768,0.01069155583779017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,1024,0.004214222232500712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,12288,0.1297395494249132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,65536,0.05343822307056851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,1024,0.006128888991143968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,768,0.041498667663998075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,2560,0.0662026670244005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,65536,0.07997155851787992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,128,0.004874666531880696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,3584,0.009233777721722921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,6144,0.009318222602208456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,2048,0.030509332815806072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,10240,0.04607466525501675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,2560,0.010991999672518836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,4096,0.00637155564294921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,64,0.00741155528359943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,6144,0.06177600224812826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,256,0.00629333323902554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,256,0.005018666800525454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,3584,0.013184888495339287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,768,0.0029582221888833572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,4096,0.0255039996571011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,6144,0.028547555208206177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,3072,0.007270221908887227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,8192,0.014931556251313953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,7168,0.01479555500878228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,16384,0.07905511061350505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,1024,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,1024,0.00907555553648207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,2560,0.005836444596449534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,128,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,65536,0.07872711287604438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,5120,0.012776888906955719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,512,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,65536,0.027040888865788777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,7168,0.015241778559154935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,1536,0.01683022247420417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,1024,0.4104986720614963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,2560,0.010062221851613786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,3072,0.0047493332790003884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,6144,0.012679110798570844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,1536,0.015810666812790763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,5120,0.008656889200210571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,65536,0.1372737752066718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,12288,0.03366577625274658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,5120,0.01648977730009291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,12288,0.04184088773197598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,4096,0.07050310903125338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,6144,0.02849510974354214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,5120,0.03390044305059645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,64,0.0057955553962124725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,65536,0.08401155471801758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,3584,0.012076444096035428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,7168,0.009683555199040307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,1024,0.007251555720965068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,768,0.00351555562681622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,7168,0.009657777845859528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,7168,0.016551110479566786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,2560,0.01613066593805949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,2048,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,6144,0.031496001614464655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,10240,0.015987555185953777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,3072,0.01588622232278188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,256,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,5120,0.05817777580685086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,6144,0.009019555317031013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,64,0.09038311243057251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,12288,0.03592355714903937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,6144,0.05253333184454176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,2560,0.004903111192915174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,256,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,16384,1.334361818101671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,12288,0.006999111009968652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,4096,0.01996888882584042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,16384,0.024939555260870192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,65536,0.0821395582622952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,1536,0.003532444437344869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,6144,0.008800889054934183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,64,0.010001777774757808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,65536,0.22779377301534018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,8192,0.0273617770936754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,3072,0.017294221454196505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,12288,0.07890222469965617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,10240,0.026382222771644592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,512,0.002528888897763358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,3584,0.027308444182078045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,768,0.00795377790927887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,256,0.015487111277050443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,16384,0.032187554571363665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,512,0.007178666690985362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,32,0.007882666256692674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,10240,0.0158933334880405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,256,0.005927111125654644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,128,0.004932444542646408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,32,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,12288,0.029044446018007066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,8192,0.025388444463411968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,2560,0.014408888088332282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,1024,0.0032026666320032547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,10240,0.05563822057512072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,8192,0.047416889005237155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,3072,0.01443377799457974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,8192,0.042872889174355396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,1024,0.021511110994550917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,3584,0.01129333343770769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,5120,0.01573422220018175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,8192,0.038231998682022095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,3584,0.03327911098798116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,8192,0.01034044474363327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,512,0.0028088889602157804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,65536,0.04403911034266154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,3584,0.03377688924471537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,256,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,1024,0.01480977733929952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,1024,0.011378666592968835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,10240,0.01794399983353085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,64,0.005647999958859549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,3072,0.009018667042255402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,512,0.01661066710948944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,768,0.0024871110088295406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,65536,0.07800800270504422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,10240,0.015803555647532146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,256,0.004568888909286923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,1536,0.006251555350091722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,64,0.013069333301650153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,65536,0.01826133330663045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,8192,0.012839111189047495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,8192,0.04484533270200094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,5120,0.03005689051416185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,1536,0.007227555745177799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,3072,0.03345866666899787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,12288,0.01275022245115704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,64,0.001826666709449556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,12288,0.035529779063330755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,256,0.015896888242827523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,5120,0.01722577710946401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,65536,0.1267671055263943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,3584,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,65536,0.04615999923812019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,3072,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,768,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,128,0.007961778177155389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,512,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,3072,0.007657777931955125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,1024,0.024248000648286607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,4096,0.00851822230550978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,6144,0.012796444197495779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,3584,0.027385777897304956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,1536,0.01035911093155543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,256,0.0170462214284473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,512,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,6144,0.013025778035322825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,1536,0.00443733317984475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,2048,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,7168,0.06274044513702393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,1024,0.006599999964237213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,4096,0.005884444548024072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,32,0.005944889038801193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,6144,0.016504888733228047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,4096,0.016192888220151264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,256,0.006365333166387346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,8192,0.16892977555592856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,5120,0.01180622229973475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,65536,0.409334209230211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,16384,0.06517066558202107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,768,0.013091555900043912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,6144,0.026541334059503343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,12288,0.03520977828237746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,64,0.005106666849719154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,65536,0.01516177753607432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,5120,0.025234666135576036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,4096,0.007605333295133378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,8192,0.02234222160445319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,7168,0.009711999860074785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,5120,0.008579555484983657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,32,0.00281866660548581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,32,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,1024,0.06082666582531399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,512,0.012085333466529846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,3072,0.026352889007992212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,32,0.001956444440616502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,3584,0.024360888534122046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,2048,0.007315555380450354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,2560,0.01015644437736935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,256,0.0032142222755485107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,768,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,3584,0.04166488846143087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,5120,0.03917422228389316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,10240,0.018989332848125033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,4096,0.021485333641370136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,2560,0.00545155546731419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,6144,0.00999644481473499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,5120,0.023967999551031325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,8192,0.01034311122364468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,2560,0.013791110780504016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,256,0.03170133299297757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,8192,0.02943199872970581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,4096,0.013120888835854001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,6144,0.008231110870838165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,8192,0.018464888135592144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,8192,0.03200622068511115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,12288,0.08882933192782932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,3072,0.00796088907453749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,6144,0.030920889642503526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,12288,0.032123555739720665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,6144,0.2957555452982585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,16384,0.024911110599835713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,128,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,7168,0.19073332680596244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,1024,0.062133332093556724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,16384,0.02269688910908169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,4096,0.020703110429975722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,512,0.013812444276279874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,7168,0.0045422220395671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,8192,0.01920088960064782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,2048,0.004179555508825514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,16384,0.04230933388074239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,4096,0.025384000606007043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,3584,0.0041911109454101985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,768,0.0027555556346972785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,10240,0.015830222103330824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,10240,0.022678222921159532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,4096,0.026112889250119526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,5120,0.7363715701633029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,1024,0.008344000412358178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,1536,0.010339555641015371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,1536,0.009500444763236577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,4096,0.044171555174721606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,7168,0.014451555079884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,7168,0.05829244189792209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,2048,0.01795288920402527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,3072,0.007213333414660559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,3584,0.022616000639067754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,7168,0.018741332822375827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,1024,0.007285333342022366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,3584,0.27167820930480957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,10240,0.008647110727098253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,1536,0.04709688822428385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,2048,0.006569777925809224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,2560,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,32,0.0018577778504954444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,6144,0.029123554627100628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,1536,0.01720533271630605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,16384,0.029560890462663438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,2560,0.010085333552625444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,3584,0.007659555309348636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,768,0.01682488951418135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,2560,0.007885333564546373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,3072,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,7168,0.04113599989149306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,128,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,3584,0.007376888559924231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,8192,0.011053333679835001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,7168,0.04173333446184794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,5120,0.004585777719815572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,768,0.01758399936887953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,2560,0.013051555388503604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,2560,0.007861333588759104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,10240,0.020972443951500788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,64,0.008719111482302347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,10240,0.020614221692085266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,1536,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,4096,0.02093066606256697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,1536,0.018244443668259513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,4096,0.01955200069480472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,2560,0.07678044504589505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,10240,0.02990488873587714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,256,0.01369511087735494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,8192,0.031367109881507024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,12288,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,3072,0.06176888942718506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,1536,0.008844444321261512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,1536,0.006299555715587404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,1536,0.012327999704413943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,1536,0.005958222266700532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,3072,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,12288,0.014803555276658801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,5120,0.004885333279768626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,6144,0.012439111040698158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,64,0.005561777700980504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,3072,0.004256000121434529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,8192,0.012030222349696688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,12288,0.007293333609898885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,32,0.005511999958091312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,2560,0.008021333151393467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,16384,0.07240355677074857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,3072,0.007205333146784041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,1024,0.01349688900841607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,1024,0.003530666646030214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,512,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,256,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,5120,0.008671110702885522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,7168,0.4431128766801622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,4096,0.2889751063452827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,128,0.016862221890025668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,6144,0.012148444851239523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,1024,0.014492443866199918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,256,0.006014222072230444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,8192,0.011928888658682505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,65536,0.06371021933025785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,7168,0.024125332633654278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,64,0.002455111179086897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,512,0.006894222150246303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,128,0.017180444465743173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,10240,0.019235556324323017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,1536,0.009871111147933537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,3072,0.02401599950260586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,10240,0.012051555845472546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,6144,0.05717421902550591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,1536,0.01456622282663981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,128,0.014870221416155497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,7168,0.010386666489972008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,2560,0.013057777451144325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,4096,0.01792711185084449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,8192,0.03513244456715054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,16384,0.14389777183532715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,4096,0.005807111246718301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,256,0.0064177778032090925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,128,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,1024,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,256,0.014781332678265043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,5120,0.009285333255926767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,64,0.005583111196756363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,768,0.018215111560291715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,8192,0.012833777401182385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,1536,0.0038702223036024305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,16384,0.07727111048168607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,65536,0.13329954942067465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,32,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,4096,0.010724444356229572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,65536,0.17834133572048613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,7168,0.010753778119881948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,5120,0.024375999967257183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,1024,0.016156444946924847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,2560,0.023032888770103455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,8192,0.010692444112565784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,128,0.011422221859296164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,3584,0.00872444444232517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,10240,3.1398711734347873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,512,0.015612444943851896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,512,0.01850222216712104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,768,0.020607999629444547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,16384,0.650203545888265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,12288,0.10877688725789388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,7168,0.027785778045654297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,65536,0.1485475566652086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,12288,1.0014106962415907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,128,0.01692800058258904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,768,0.011504888534545898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,4096,0.013584888643688627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,10240,0.054745776785744556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,7168,0.016186666157510545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,32,0.002198222196764416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,128,0.0027582223216692605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,128,0.00886933339966668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,768,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,1024,0.00795199970404307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,16384,0.034685333569844566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,1024,0.007234666910436418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,10240,0.029194666279686823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,8192,0.1375653346379598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,65536,0.13286133607228598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,12288,0.02365422248840332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,65536,0.39338310559590656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,128,0.019696000549528334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,256,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,768,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,5120,0.0188417782386144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,3584,0.009425777528021071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,7168,0.02750844425625271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,2048,0.014451555079884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,3072,0.0075351107451650835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,8192,0.036128888527552284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,1024,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,5120,0.008298666940795051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,2048,0.008972444468074376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,2560,0.012411555482281579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,16384,0.015466666883892484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,12288,0.07606755362616645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,65536,0.06562666760550605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,16384,0.3664168781704373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,3584,0.03871999846564399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,2048,0.007112888826264276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,256,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,5120,0.01034755590889189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,3584,0.018650665879249573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,3072,0.013410666750537025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,4096,0.01147377739350001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,768,0.008270222279760573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,2560,0.007262222468852997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,3072,0.07363022036022611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,768,0.0031679999083280563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,2048,0.03276088833808899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,6144,0.24246843655904135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,2560,0.006641777853171031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,10240,0.02230577833122677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,16384,1.1875875261094835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,16384,0.3651973406473796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,2048,0.016379555066426594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,512,0.0037964445849259696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,6144,0.0417795545525021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,3584,0.016184889607959323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,1024,0.026621333426899377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,768,0.002502222235004107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,4096,0.008294222255547842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,7168,0.011804444094498953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,64,0.0052328887912962176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,768,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,12288,0.019031110737058852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,5120,0.007669332954618666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,4096,0.01903199983967675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,3072,0.009753777749008602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,65536,0.3106711175706652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,768,0.011063999599880643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,7168,0.011524444652928246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,16384,0.03165955675972833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,6144,0.005991999887757831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,2560,0.08677599827448527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,4096,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,12288,0.012714666624863943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,8192,0.011326221956147088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,16384,0.07644444704055786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,3584,0.008486222061845992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,3072,0.00961511085430781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,2048,0.00424888895617591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,2048,0.01090844472249349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,128,0.002163555473089218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,16384,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,256,0.006609777609507243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,768,0.005456888841258154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,6144,0.004915555732117759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,10240,0.030395554171668157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,2560,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,3584,0.01716444392999013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,12288,0.05915821923149956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,128,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,7168,0.018193778064515855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,8192,0.010682666467295753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,10240,0.011400889191362592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,16384,0.01807466646035512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,128,0.005610666755172942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,768,0.01497688889503479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,32,0.002557333351837264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,16384,0.21987111038631865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,4096,0.05432800120777554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,65536,0.08268621895048353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,256,0.005681777579916849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,65536,0.07267110877566867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,2560,0.006590222318967183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,3584,0.024849777420361836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,65536,0.07233066691292657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,32,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,10240,0.01555377741654714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,8192,0.08802844418419732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,2048,0.006301333506902059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,256,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,6144,0.39947999848259824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,2560,0.009415999882751042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,1024,0.004212444441186057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,256,0.006976889239417181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,6144,0.009367110828558603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,2048,0.006278222219811545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,2048,0.00902133352226681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,65536,1.0234301884969075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,7168,0.037288000186284385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,1024,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,10240,0.04457155532307095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,1536,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,4096,0.007962666451931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,768,0.03028622269630432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,256,0.0255422227912479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,8192,0.024484443995687697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,7168,0.018894222047593858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,4096,0.025291555457644995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,1536,0.023543111152119104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,128,0.0020311110549502904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,10240,0.020270221763186984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,10240,0.14346488316853842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,10240,0.01924533314175076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,1024,0.01959911154376136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,6144,0.06752622127532959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,8192,0.00869066682126787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,7168,0.02739022175470988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,12288,0.042318221595552236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,3072,0.00797155582242542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,10240,0.03397066725624932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,6144,0.00998933364947637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,7168,0.03285155693689982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,10240,0.012089778151777057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,512,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,8192,0.10752533541785346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,4096,0.06470844480726454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,16384,0.030812445614073012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,1024,0.0034888887570963967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,7168,0.018213333355055917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,2048,0.01112711098459032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,12288,0.4148453341590033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,128,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,7168,0.015106666419241162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,3584,0.029566221767001685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,5120,0.011384888655609555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,3072,0.008637333081828224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,65536,0.059600002235836454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,12288,0.020792000823550753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,3584,0.013160000244776407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,768,0.005626666463083691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,2048,0.018922666708628338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,256,0.01443199978934394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,32,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,1536,0.007642666498819987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,12288,0.017758222089873422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,7168,0.010799999866220685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,256,0.09532088703579372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,2048,0.008547555241319869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,1024,0.005997333261701796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,7168,0.01593777702914344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,256,0.0025271111064487034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,6144,0.049666666322284274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,7168,0.03693866729736328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,1024,0.016582222448454965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,2048,0.006993778049945831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,128,0.0032044444233179092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,8192,0.007675555845101674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,3584,0.007334222396214803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,32,0.00552977787123786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,2048,0.015481778317027621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,32,0.0022026666750510535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,8192,0.05229777759975857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,6144,0.19308800167507598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,16384,0.3627155621846517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,10240,0.01738933391041226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,6144,0.09555111328760783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,64,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,2560,0.012551111479600271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,10240,0.10996177461412217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,768,0.010371555884679159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,2048,0.007653333246707916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,512,0.002536888958679305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,65536,0.07397511270311144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,2048,0.007408888803588019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,8192,0.028742220666673448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,10240,0.019595555133289762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,2048,0.004210666649871402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,10240,0.009494221872753566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,8192,0.12166933218638103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,65536,0.06895466645558675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,5120,0.019647111495335896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,8192,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,6144,0.03360088997417026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,3584,0.010785777535703448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,65536,0.15228266186184353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,64,0.0021804444905784395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,7168,0.21707822216881645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,512,0.01311466677321328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,2048,0.09410844246546428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,65536,0.07665599717034234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,5120,0.021287999219364587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,1024,0.008375999828179678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,768,0.015187554889255099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,2560,0.0069004446268081665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,65536,0.6820186509026421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,128,0.01205244412024816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,32,0.002483555633160803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,16384,0.07740977737638685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,128,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,10240,0.04787288771735298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,6144,0.05988444222344292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,128,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,5120,0.013105777402718862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,6144,0.026543999711672466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,3584,0.007607111500369177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,8192,1.246996455722385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,512,0.009333333207501305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,65536,0.8198595576816134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,1024,0.014118222726715935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,1024,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,5120,0.008293333152929941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,1536,0.03467199868626065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,12288,0.014118222726715935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,6144,0.008940444224410588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,32,0.003866666720973121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,2048,0.009422221945391761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,7168,0.009688888986905416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,2560,0.004889777965015835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,10240,0.005199111170238919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,7168,0.03552800085809495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,1536,0.007328888608349695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,6144,0.014783110883500842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,4096,0.030174222257402208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,4096,0.016571554872724745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,1024,0.013140444126394061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,64,0.002187555655837059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,3072,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,128,0.0018648888087934917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,128,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,4096,0.14035111003451878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,512,0.00795999997191959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,6144,0.01956355571746826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,8192,0.02144533395767212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,256,0.002159111201763153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,5120,0.014305777019924588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,4096,0.008079110748238033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,2048,0.04078400135040283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,512,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,10240,0.07016533613204956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,256,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,2560,0.010083555347389644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,5120,0.009256000320116678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,2048,0.02718755602836609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,256,0.002469333302643564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,4096,0.01111822244193819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,128,0.009398221969604492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,65536,0.06785066922505696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,7168,0.04740711053212484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,4096,0.09680977794859146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,256,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,16384,0.016898666818936665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,12288,0.05236533284187317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,4096,0.02193244463867611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,8192,0.018570666511853535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,6144,0.023911111884646948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,4096,0.019920888874265883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,8192,0.022367111510700647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,768,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,16384,0.012808000048001608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,2560,0.013360000318951078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,6144,0.015912888778580558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,12288,0.011983999775515662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,4096,0.020078221956888836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,5120,0.009351111120647853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,2560,0.009104889300134447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,128,0.005056888692908817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,1536,0.004563555535342959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,6144,0.009303111169073317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,16384,0.028677334388097126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,2048,0.011500444677140979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,256,0.007655999726719326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,3072,0.03499911228815714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,10240,0.04784799946679009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,32,0.0024826667375034755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,10240,0.0436097780863444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,65536,0.11310666137271458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,64,0.008276444342401292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,5120,0.010388444695207808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,512,0.013408888545301227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,1024,0.0034799998005231223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,128,0.005208000126812193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,512,0.0042275554604000514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,10240,0.052834666437572904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,1536,0.021822222405009802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,6144,0.012647999657524956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,2560,0.0063795554969045855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,128,0.0162773331006368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,3584,0.02516799999607934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,3584,0.014135999812020196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,6144,0.009360888765917884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,65536,0.12966133488549128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,2560,0.043149332205454506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,5120,0.42556354734632706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,3072,0.007253333098358578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,256,0.005958222266700532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,5120,0.027365333504146997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,2048,0.004978666702906291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,768,0.005619555711746216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,2048,0.0127688886390792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,1536,0.006274666637182236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,3072,0.014478221535682678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,1536,0.006232000059551663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,2048,0.01479555500878228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,16384,0.00562755556570159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,2048,0.013767999907334646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,65536,0.22563910484313965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,8192,0.018381334013409086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,12288,0.02481777800454034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,2560,0.014105778601434497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,512,0.01129422254032559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,5120,0.03517244590653314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,65536,0.12457956208123101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,2048,0.02146133283774058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,2560,0.00489688871635331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,10240,0.0306311117278205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,2560,0.013415999710559845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,1536,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,3584,0.017853332890404593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,5120,0.03887466589609782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,12288,0.030074665943781536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,2560,0.00692622239391009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,512,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,4096,0.01683022247420417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,65536,0.044512000348832875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,8192,0.03793511125776503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,128,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,4096,0.007905777957704332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,65536,0.18629422452714708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,3584,0.09882933563656277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,128,0.006537777682145436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,1536,0.004654222064548069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,1024,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,16384,0.0365973346763187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,2048,0.01717688971095615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,4096,0.0163128889269299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,32,0.0018151110659042995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,12288,0.024725332856178284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,12288,0.039140443007151283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,16384,0.05184266633457608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,10240,0.06957511107126872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,256,0.014562666416168213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,256,0.02910755409134759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,768,0.029620445436901514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,1536,0.003532444437344869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,4096,0.01614755557643043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,6144,0.011786667009194693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,1024,0.0058017778727743365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,3584,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,3584,0.010594666832023196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,4096,0.008192000289758047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,5120,0.013945778210957846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,2560,0.006364444477690591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,128,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,7168,0.011876444849703046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,768,0.08490222030215794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,10240,0.05610577927695381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,5120,0.009180444810125563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,8192,0.045327110422982105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,4096,0.00909511082702213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,8192,0.021105777886178758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,768,0.009272000028027428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,3584,0.00793422261873881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,2560,0.010346666806273991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,10240,0.02199999988079071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,10240,0.007956444389290279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,10240,0.04229777720239428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,12288,0.05930578046374851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,512,0.0058711109062035876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,4096,0.03128444486194187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,512,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,2048,0.004600888739029567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,768,0.002757333426011933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,1536,0.004567111117972268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,64,0.002128888956374592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,2048,0.03326844506793552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,1024,0.011728000309732227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,128,0.004902222090297275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,3584,0.007582222421964009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,512,0.013433777623706393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,3072,0.0232631117105484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,768,0.07649422354168363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,1536,0.006196444647179709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,5120,0.00906577789121204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,65536,0.051706665092044406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,4096,0.008921778036488427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,1024,0.01482488877243466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,12288,0.029065777858098347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,16384,0.04452800088458591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,4096,0.011481777661376528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,7168,0.012151999606026543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,256,0.005274666680230035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,512,0.002692444456948174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,8192,0.015328889091809591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,10240,0.009092444347010719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,6144,0.021707556313938562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,2048,0.0075048887067370946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,3584,0.010046222143703038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,12288,0.05016355713208517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,5120,0.008681777450773451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,10240,0.023014222582181294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,65536,0.07984889215893216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,4096,0.02211911148495144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,1536,0.01756000022093455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,4096,0.014867555763986377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,1536,0.013427555561065674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,1536,0.02856711215443081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,256,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,1024,0.013765333427323235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,2560,0.01650311052799225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,65536,0.07744444741143121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,4096,0.06180177794562446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,5120,0.027208000421524048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,1536,0.006170666466156642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,3584,0.00867555538813273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,3072,0.014845333165592618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,5120,0.02677511175473531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,65536,0.18963466750250924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,8192,0.27239465713500977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,7168,0.023383999864260357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,3072,0.012087110843923358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,6144,0.012846222354306115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,4096,0.011560888753996955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,65536,0.12283821900685628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,512,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,7168,0.01720977822939555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,128,0.005280888742870755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,1024,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,65536,0.10576888587739731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,12288,0.033407999409569636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,128,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,10240,0.019334221879641216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,512,0.021720000439220007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,7168,0.056333336565229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,512,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,65536,0.07782577806048922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,7168,0.013076444466908773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,256,0.014113777213626437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,7168,0.00979466653532452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,10240,0.01942311061753167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,8192,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,64,0.006950221955776215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,5120,0.01718844473361969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,2048,0.019104889697498746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,2048,0.006559999866618051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,512,0.005942222144868638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,768,0.014577777849303352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,12288,0.019243554936514962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,3072,0.015923556354310777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,8192,0.028567999601364136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,2560,0.005020444591840108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,7168,0.00979200005531311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,65536,0.06059110826916165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,5120,0.03608622153600057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,3072,0.009533333281675974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,10240,0.011733333269755045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,8192,0.03014844324853685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,3584,0.016757332616382174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,512,0.05728889173931546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,128,0.006342222293217977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,3072,0.015899555550681222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,5120,0.015618667006492615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,16384,0.015167999598715039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,5120,0.05064444409476387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,512,0.005607999861240387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,256,0.0022151112142536375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,1024,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,64,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,2048,0.01178311142656538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,256,0.004927111168702443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,5120,0.12130488289727105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,256,0.004910222358173794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,5120,0.03265955713060167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,4096,0.007629333270920648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,7168,0.01072177787621816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,3584,0.017850667238235474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,4096,0.022054221895005967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,5120,0.015544889701737298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,128,0.005888888819350137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,5120,0.07342400153477986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,10240,0.03239733311865065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,16384,0.018820444742838543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,4096,0.009983110758993361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,256,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,64,0.002485333424475458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,6144,0.023785778217845496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,768,0.007010666860474481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,3584,0.010336000058386061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,2560,0.032912890116373696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,1536,0.015582222077581616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,2048,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,6144,0.02604177759753333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,768,0.005544888890451855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,128,0.005111111121045219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,64,0.0017884444031450485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,256,0.031762666172451444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,512,0.002551111082235972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,3584,0.018584888842370775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,12288,0.18462488386366102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,6144,0.013105777402718862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,65536,0.13407556215922037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,10240,0.01906399925549825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,16384,0.052912000152799815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,2560,0.014107555150985718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,8192,0.007328888608349695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,6144,0.00999466660949919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,2560,0.007730666961934831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,6144,0.009365333451165093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,2048,0.021253334151373968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,7168,0.01566844516330295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,6144,0.009342222577995723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,1024,0.015287111202875773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,64,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,512,0.016568000117937725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,128,0.0022737777067555324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,8192,0.010701333483060202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,8192,0.008771555291281806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,16384,0.18872800138261583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,8192,0.01993333299954732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,8192,0.025739555557568867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,768,0.01718755563100179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,7168,0.04043466515011258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,1536,0.01149511088927587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,1536,0.017198221551047433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,7168,0.015150222513410779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,256,0.004936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,5120,0.012758221891191272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,6144,0.01886755559179518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,128,0.014160000615649752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,4096,0.007649777664078607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,128,0.005199999858935674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,3584,0.005604444278611078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,768,0.005924444645643234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,16384,0.02341777748531765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,5120,0.03360266817940606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,8192,0.015466666883892484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,3584,0.007344889144102733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,256,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,128,0.004922666483455234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,1024,0.0035253332720862497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,16384,0.06746488809585571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,4096,0.011348444554540845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,2048,0.026878222823143005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,16384,0.10155022144317627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,3072,0.08909155262841119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,3584,0.008663111262851292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,2048,0.020414221617910597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,12288,0.010354667074150508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,768,0.004927999857399198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,6144,0.015457777513398064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,8192,0.044551998376846313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,256,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,256,0.014509333504570855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,768,0.01721155477894677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,128,0.014091556270917257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,16384,0.015465777781274585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,10240,0.023316444622145757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,2048,0.006595555692911148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,10240,0.005609777652555042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,768,0.007315555380450354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,1536,0.07402400175730388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,3584,0.015469332536061605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,512,0.002480888946188821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,7168,0.009649777577983009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,10240,0.024015110399987962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,6144,0.026534222894244727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,8192,0.06711110803816053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,65536,0.10116355286704169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,4096,0.10902488893932766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,2048,0.017469333277808297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,768,0.005190222213665645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,512,0.012415111064910889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,128,0.0018044445249769424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,3584,0.008740444150235917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,128,0.013075555364290873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,2560,0.015799111790127225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,65536,0.08110933171378241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,8192,0.005958222266700532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,32,0.0019128889673286015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,8192,0.011008000208271874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,512,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,5120,0.01240799989965227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,1024,0.009343110852771336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,2560,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,65536,0.05271822214126587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,7168,0.12780355082617864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,65536,0.09419200155470107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,12288,0.011708444191349877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,7168,0.017830222845077515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,2048,0.02351733379893833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,16384,0.10198933548397487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,16384,0.015124445160230001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,7168,0.028135998381508723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,65536,0.11812178293863933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,8192,0.012075555821259817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,1024,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,3584,0.021087111698256597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,256,0.00453955555955569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,16384,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,2048,0.006254222244024277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,12288,0.5681226518419054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,32,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,512,0.002838222309947014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,7168,0.024137778414620295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,16384,0.0225911107328203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,3072,0.012335110869672565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,12288,0.03401155604256524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,64,0.01036888857682546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,65536,0.08958399958080715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,10240,0.028343998723559912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,3584,0.008392888638708327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,6144,0.0110880004035102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,5120,0.012141333685980903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,256,0.005881777654091517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,1536,0.013222222526868185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,6144,0.01923377811908722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,8192,0.08338489135106404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,5120,0.019280888968043856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,4096,0.008636444807052612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,3584,0.005970666805903117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,4096,0.010716444088353051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,65536,0.04753155509630839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,7168,0.4165288872188992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,16384,0.045915553967158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,2048,0.009007111191749573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,3584,0.008229333493444655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,8192,0.011381333072980246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,2048,0.02233422299226125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,2560,0.019137778215938143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,3072,0.009002666506502364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,6144,0.009708444277445475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,2560,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,10240,0.01204711116022534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,768,0.006976000136799282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,1536,0.006011555592219035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,1024,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,8192,0.024455999334653217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,4096,0.07943911022610135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,10240,0.010966222319338055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,10240,0.03313333458370633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,8192,0.008632889224423302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,5120,0.005561777700980504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,2048,0.005154666801293691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,1024,0.00796977761718962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,2560,0.015127110812399121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,768,0.0028391112056043413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,32,0.001810666587617662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,65536,0.16196621788872614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,256,0.009787555370065901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,2560,0.007056000332037608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,3584,0.17160977257622612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,10240,0.015565334094895257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,2048,0.004353777815898259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,2560,0.007355555891990662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,4096,0.006663110935025745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,6144,0.05163022213511997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,32,0.0028986665937635633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,8192,0.05441688828998142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,128,0.003560888684458203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,3584,0.007616889145639207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,768,0.01482311056719886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,768,0.0414959987004598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,768,0.01824711097611321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,6144,0.010856889188289642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,16384,0.028592000404993694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,128,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,3072,0.08749688996209039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,1536,0.029986666308508977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,1536,0.0069653333889113525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,3072,0.007646222081449296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,768,0.00452977791428566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,128,0.014818666709793938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,1024,0.003212444484233856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,768,0.03187111020088196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,12288,0.049433777729670204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,256,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,5120,0.010457778142558204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,6144,0.013462222284740873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,8192,0.01517511076397366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,10240,0.028326223293940227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,2560,0.007318221860461765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,4096,0.011055111057228513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,16384,0.019271110494931538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,16384,0.03354933195643955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,64,0.0035235554807715942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,3584,0.011687999798191918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,4096,0.00792622235086229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,256,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,5120,0.008621333373917473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,10240,0.08786133262846206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,7168,0.04259377717971802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,1536,0.005931555396980709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,3584,0.004238222208287981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,3072,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,10240,0.5723830858866373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,1024,0.0035217776894569397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,7168,0.01706133286158244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,65536,0.3648071024152968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,10240,0.022740445203251306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,2048,0.009311999711725447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,1536,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,512,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,128,0.014447111222479077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,6144,0.012409778104888068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,12288,0.02682222260369195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,1024,0.009371555513805812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,7168,0.020273778173658583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,7168,0.030380444394217596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,10240,0.03051999873585171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,5120,0.040417777167426214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,2560,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,3072,0.021550221575631037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,2048,0.006619555668698416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,2048,0.006343110981914733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,10240,0.057674666245778404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,12288,0.01501066651609209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,3072,0.00463377767139011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,6144,0.015671110815472074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,2560,0.008984000318580205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,3072,0.03158311049143473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,6144,0.009031111167536842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,1024,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,64,0.0028231110837724474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,1536,0.029262221521801416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,65536,0.07933688826031156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,6144,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,64,0.0042071110672420925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,3584,0.015832889411184523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,4096,0.04507466819551256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,10240,0.026188444760110643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,4096,0.03532444437344869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,10240,0.405079099867079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,1536,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,10240,0.02646044393380483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,2048,0.010515555739402771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,16384,0.0196426659822464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,10240,0.8005244466993543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,1024,0.012450666891203986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,16384,0.22907998826768664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,16384,0.026875555515289307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,768,0.006589333216349284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,768,0.04150044586923387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,256,0.011722666521867117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,1024,0.007713778151406183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,12288,0.02306844459639655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,1536,0.005041777673694822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,5120,0.041542222102483116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,256,0.0022693334354294673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,3072,0.05135644475618998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,5120,0.0329022208849589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,512,0.012774222426944308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,256,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,16384,0.011300444602966309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,256,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,2560,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,2048,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,12288,0.013807111316257052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,12288,0.020999999509917364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,128,0.005019555489222209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,3584,0.13416266441345215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,4096,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,256,0.011047999891969891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,2560,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,12288,0.052834666437572904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,1024,0.06334577666388617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,512,0.030208001534144085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,128,0.005203555441564984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,128,0.013400889105266996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,1536,0.0038408889538711975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,12288,0.0052986666560173035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,3072,0.014312000738249885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,256,0.005208000126812193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,512,0.003170666595300039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,3584,0.026530666483773127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,2048,0.02816533380084568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,1536,0.04864533411131965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,12288,0.07126310798856947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,2560,0.030179556873109605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,16384,0.06443911128573947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,10240,0.022321777211295232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,128,0.0028773333049482773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,768,0.011899555722872416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,128,0.01442399952146742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,65536,0.8370035489400228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,1024,0.029215110672844782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,256,0.015822221835454304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,1536,0.016506666938463848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,10240,0.02581688927279578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,8192,0.03999555442068312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,6144,0.029453333881166246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,2560,0.06090044313006931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,16384,0.038113776180479265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,12288,0.024653333756658766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,1024,0.026664000418451097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,10240,0.020920889245139226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,5120,0.1229777733484904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,16384,0.02882755464977688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,10240,0.06773244672351413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,128,0.015122666954994202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,1536,0.013801777528391944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,3072,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,7168,0.11092355516221787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,256,0.015200000670221118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,16384,0.17145066791110566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,768,0.007135999699433644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,8192,0.046783109505971275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,1536,0.00831911133395301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,7168,0.21021600564320883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,2560,0.014367111855083041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,2048,0.004611555486917496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,16384,0.027300443914201524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,128,0.0021848889688650766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,128,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,6144,0.009447111023796929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,1536,0.024757333927684363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,768,0.007000889215204451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,6144,0.10458399852116902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,3584,0.008811555802822113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,1024,0.026139555705918208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,1024,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,7168,0.029281778468026057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,256,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,12288,0.06349777513080172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,64,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,2560,0.007655999726719326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,12288,0.023883556326230366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,65536,0.0771920018725925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,12288,0.5294515291849772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,3584,0.010061333576838175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,4096,0.00795111142926746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,10240,0.020395555429988436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,7168,0.014933332800865173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,10240,0.012072000238630505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,12288,0.03844088978237576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,6144,0.02111377815405528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,10240,0.011511999699804517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,65536,0.11556888951195611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,12288,0.02973244587580363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,2560,0.016551999582184684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,5120,0.00831822223133511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,2560,0.0069191112286514705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,128,0.006015111174848344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,4096,0.03015111221207513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,16384,0.04589777853753832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,7168,0.024636444118287828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,256,0.004955555415815777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,10240,0.018416888184017606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,512,0.002467555511328909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,128,0.001975111042459806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,7168,0.04141599933306376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,3072,0.0041626666982968645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,6144,0.015840889679061044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,64,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,768,0.011172444456153445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,3072,0.044028444422615894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,65536,0.05552177627881368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,512,0.0027760000278552375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,512,0.010021333065297868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,128,0.005205333232879639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,7168,0.012802666260136498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,8192,0.05707910988065931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,1536,0.02482844392458598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,4096,0.04053066505326165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,128,0.0021946666141351065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,32,0.011524444652928246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,7168,0.009687110781669617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,512,0.002552888873550627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,3072,0.007313778003056844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,3072,0.01789688898457421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,16384,0.18991733921898735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,7168,0.005524444497293896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,4096,0.011543999943468304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,16384,0.02574488851759169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,10240,0.4321137799157037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,4096,0.018926221463415358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,3584,0.025504888759719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,65536,0.04493155413203769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,8192,0.04827910992834303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,5120,0.016857778032620747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,1024,0.012088889049159156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,2048,0.008070222205585903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,1024,0.006219555520349079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,512,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,4096,0.010395555860466426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,2560,0.007608888877762689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,1024,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,768,0.01481244464715322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,128,0.023437332775857713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,6144,0.01722666621208191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,5120,0.05473955472310384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,16384,0.020968000094095867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,12288,0.09550844298468696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,4096,0.13014400005340576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,4096,0.11842755476633708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,128,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,3072,0.013597333596812354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,6144,0.020911999874644812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,7168,0.035233779086007014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,64,0.006607999818192587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,256,0.006280888699822956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,2560,0.0070017774899800616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,16384,0.025069332785076563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,768,0.003999111139112049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,2560,0.05520711011356778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,12288,0.058225777414109975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,3584,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,1024,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,3584,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,12288,0.019164444671736825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,1024,0.025394666526052687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,4096,0.014602666099866232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,3072,0.010703110860453712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,1024,0.0028640000770489373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,768,0.016551999582184684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,8192,0.03049600124359131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,7168,0.007663110891977946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,1024,0.014819555812411837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,10240,0.15881778134240046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,128,0.005199111170238919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,7168,0.010710222025712332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,12288,0.012726222475369772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,8192,0.03679910964436001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,256,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,4096,0.03914577762285868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,3584,0.005583999885453119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,2048,0.014436443646748861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,7168,0.02698488864633772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,7168,0.027200889256265428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,32,0.00216977774269051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,5120,0.00453955555955569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,16384,0.2600488927629259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,4096,0.012079999678664737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,12288,0.04202399982346428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,10240,0.016843555702103507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,2560,0.015462223026487561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,128,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,64,0.0018622221218215094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,1024,0.005631111148330901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,10240,0.02409422232045068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,16384,0.015509333875444202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,1024,0.006011555592219035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,256,0.009558222360081142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,768,0.07268177800708346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,3584,0.007771555748250749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,12288,0.02714666724205017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,1024,0.013855111267831592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,3584,0.060311112138960093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,4096,0.02941155433654785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,6144,0.01404711107412974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,768,0.03839377893341912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,12288,0.06062755319807264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,6144,0.009296889106432596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,12288,0.019734222027990553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,16384,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,10240,0.02604444490538703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,7168,0.06397866540484957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,7168,0.020757334099875558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,12288,0.020587555236286588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,65536,0.0759084423383077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,2048,0.006263999889294307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,2560,0.0069191112286514705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,512,0.013478221992651621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,16384,0.02417244513829549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,65536,1.2825280295477974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,16384,0.021334222621387903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,7168,0.10089511341518825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,768,0.003834666477309333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,4096,0.004315555509593752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,6144,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,65536,2.738576889038086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,128,0.011746666497654386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,512,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,2560,0.007666666474607255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,2048,0.007432888779375289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,8192,0.011420444481902652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,64,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,768,0.01755822201569875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,2560,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,12288,0.03700000047683716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,2560,0.006712889091836081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,1024,0.00589155571328269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,128,0.013464888764752282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,16384,0.02831733226776123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,32,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,3072,0.03729333480199178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,768,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,3072,0.054233776198493115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,32,0.0025635556214385563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,12288,0.06184177928500705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,2560,0.006923555499977536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,10240,0.012505778008037143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,5120,0.008652444514963362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,10240,0.01531733406914605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,128,0.009830222361617619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,128,0.009655111365848118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,65536,0.14765955342186823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,512,0.008992888861232335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,8192,0.015928000211715698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,65536,1.1122400495741103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,7168,0.013285333083735572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,16384,0.2194195588429769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,1024,0.014492443866199918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,6144,0.0559448864724901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,12288,0.1530159976747301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,64,0.005927111125654644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,65536,0.057927111784617104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,2048,0.022053332792388067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,65536,0.08131911357243855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,2560,0.05915021896362305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,65536,0.5321875678168403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,16384,0.05814933114581638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,1536,0.006552000012662675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,3584,0.01088177743885252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,3584,0.0075706665714581805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,65536,0.0669306649102105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,4096,0.007884444461928474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,3072,0.01817599932352702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,2048,0.01127199994193183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,7168,0.004941333499219683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,3584,0.009332444104883406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,3072,0.013481777575280933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,3072,0.016187555260128446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,1536,0.0817084444893731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,768,0.061905774805280894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,1536,0.0035439998739295532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,5120,0.023990222149425085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,32,0.01754844519827101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,16384,0.06499289141760932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,7168,0.01850666602452596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,7168,0.009897777603732215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,1536,0.0038497779104444715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,65536,0.21088622676001656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,65536,0.21893066830105254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,2048,0.006924444602595435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,7168,0.0392542216512892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,3584,0.13559199704064265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,7168,0.02102222210831112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,768,0.0029075555503368378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,1024,0.004853333450025982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,2560,0.02236977716286977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,8192,0.01351199961370892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,2048,0.005525333185990651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,3584,0.15585066212548151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,16384,10.78149159749349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,256,0.02442399991883172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,8192,0.020247111717859905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,8192,0.20877956019507515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,12288,0.019484443797005545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,2560,0.013820444544156393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,16384,0.027239110734727647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,10240,0.045384890503353544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,768,0.00535022219022115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,256,0.006252444452709622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,512,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,2048,0.011085333095656501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,7168,0.02180444366402096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,3584,0.016872000363137987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,256,0.013225778109497495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,256,0.012738666600651212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,5120,0.09442755248811509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,32,0.005953777581453323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,2048,0.006232888748248418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,10240,0.018584888842370775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,128,0.014486221803559197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,8192,0.010140444669458602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,12288,0.030621333254708186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,7168,0.012025777664449481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,256,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,128,0.005249777601824866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,1536,0.02235200007756551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,5120,0.006264888991912206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,1024,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,512,0.003176888864901331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,16384,0.10967199669943915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,65536,0.012538666526476542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,16384,0.02371999952528212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,5120,0.008661333057615492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,2048,0.009568889107969072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,128,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,128,0.015146666102939181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,5120,0.05739111370510525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,3072,0.022985777921146814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,32,0.003900444342030419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,5120,0.0107004443804423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,32,0.002159111201763153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,16384,0.024745777249336243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,2560,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,2048,0.012083555261294046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,7168,0.009701333112186855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,768,0.014688889185587565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,6144,0.009332444104883406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,256,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,5120,0.02629866699377696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,2560,0.013797333670987023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,64,0.0021013333979580137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,128,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,6144,0.02513244416978624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,12288,0.020629333125220407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,2048,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,512,0.008833777573373582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,3072,0.015402666396564908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,12288,0.01754488878779941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,1536,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,2048,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,6144,0.02946399980121189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,768,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,8192,0.12341777483622234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,1536,0.00886311133702596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,8192,0.013057777451144325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,8192,0.3927191098531087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,65536,0.019263111882739596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,2048,0.08012711339526706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,3584,0.01587733295228746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,128,0.0031395554542541504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,7168,0.043763554758495755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,12288,0.11078133185704549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,512,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,64,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,10240,0.00851644410027398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,3584,0.014811555544535318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,768,0.02103288968404134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,65536,0.29922578069898814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,12288,0.13113421863979763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,6144,3.9813109503852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,6144,0.0385519994629754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,6144,0.022295110755496558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,64,0.002504000026318762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,6144,0.02788355615403917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,768,0.01889155473974016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,1536,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,10240,0.02126933303144243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,64,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,1024,0.01923377811908722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,512,0.005917333480384614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,64,0.002512000087234709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,4096,0.15262666013505724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,16384,0.02402755618095398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,3584,0.013817778064144982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,12288,0.008419555922349295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,128,0.002011555557449659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,4096,0.021677333447668288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,1536,0.02700177828470866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,3584,0.010348444183667501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,3072,0.012641777594884237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,6144,0.008935111264387766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,768,0.017458667357762653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,64,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,6144,0.4433991114298503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,6144,0.018815999229749043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,64,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,10240,0.10922666390736897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,10240,0.021087111698256597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,10240,0.03667111198107401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,10240,0.02500711050298479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,512,0.0041928887367248535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,32,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,32,0.005537777725193236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,2560,0.040017777019076876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,10240,0.025040000677108765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,4096,0.005609777652555042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,256,0.009343999955389235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,2048,0.00666933341158761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,3584,0.3797315491570367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,1024,0.006326222171386083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,64,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,3584,0.011046222514576383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,1024,0.013408000270525614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,128,0.007265778051482306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,1024,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,8192,0.1985306607352363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,256,0.0022071111533376905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,16384,0.05145333210627238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,8192,0.01756533318095737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,7168,0.018567111757066514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,7168,0.09666044182247585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,3072,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,512,0.018180444836616516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,64,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,3584,0.02866577770974901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,7168,0.017743999759356182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,3584,0.0058853332367208265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,2048,0.006324444380071428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,12288,0.17516177230411104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,8192,0.029976000388463337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,512,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,1024,0.01682400041156345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,32,0.004895110925038655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,7168,0.022222222553359136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,6144,0.021586666504542034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,2048,0.10321599907345241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,16384,0.03292266527811686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,768,0.014785778191354541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,4096,0.007599111232492659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,256,0.004873777843183941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,4096,0.008346666892369589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,10240,0.02545333405335744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,8192,0.0052586665583981415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,3072,0.006909333169460297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,5120,0.0107004443804423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,1536,0.01516177753607432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,1024,0.025627555118666753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,32,0.00184888889392217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,2560,0.03258666727277968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,6144,0.02387111054526435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,6144,0.009663110805882348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,2560,0.011482666763994427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,1024,0.009294222626421187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,12288,0.02476799984773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,256,0.007461333440409766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,3072,0.007947555846638149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,64,0.002185777864522404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,4096,0.011376000112957425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,3072,0.007344889144102733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,2560,0.006849777781301075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,65536,0.055148445897632174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,2560,0.01686844395266639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,8192,0.020577778418858845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,2560,0.01824711097611321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,65536,1.3662524753146703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,7168,0.14424888292948404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,128,0.0021377777059872947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,4096,0.01920977731545766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,768,0.0031866667171319327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,2048,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,16384,0.03156088789304098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,16384,0.23407554626464844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,4096,0.010560000108347999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,12288,0.017316444052590262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,7168,0.01573866605758667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,65536,0.07354933023452759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,32,0.003516444315512975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,2560,0.030685332086351182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,12288,0.03590222199757894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,1024,0.03886311252911886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,2560,0.07393866777420044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,256,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,6144,0.009678222239017487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,12288,0.0125591109196345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,5120,0.026910222238964502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,512,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,3584,0.007621333003044128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,3072,0.022660444180170696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,768,0.007303111255168915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,256,0.00490133340160052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,12288,0.03122666809293959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,12288,0.025981333520677354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,768,0.008977777428097194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,1536,0.021724444296624925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,5120,0.014624888698259989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,6144,0.009707555174827576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,10240,0.019824888971116807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,6144,0.005608888963858287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,2560,0.04765244325002035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,128,0.013470222552617392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,2560,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,65536,0.11485777960883246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,4096,0.02436888880199856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,2048,0.006291555447710886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,12288,0.057052446736229785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,768,0.017250667015711468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,256,0.0032248888164758682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,128,0.006255110932721033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,12288,0.036413331826527916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,1536,0.01795377830664317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,8192,0.010318222145239512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,128,0.013755555782053204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,2560,0.007381333245171442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,12288,0.05130755570199755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,256,0.006239999913507038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,1024,0.0028373334142896864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,65536,0.0817013316684299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,1536,0.022649778260125056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,512,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,6144,0.016238222519556682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,65536,1.6371581819322374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,3584,0.01681422193845113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,10240,0.01581333412064446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,1536,0.006218666831652324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,32,0.00758755538198683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,3072,0.028063111835055884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,4096,0.0266311119000117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,12288,0.023788443870014612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,768,0.018995554910765756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,2048,0.006953777538405524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,3584,0.008007110820876228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,8192,0.02615822189384037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,64,0.0018702221827374564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,65536,0.05583733320236206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,3584,0.6146444214714898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,32,0.007272889216740926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,2048,0.006959111326270633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,768,0.005950222412745158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,64,0.002135111019015312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,12288,0.02164177762137519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,1024,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,6144,0.04969777663548788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,10240,1.7310462527804906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,1536,0.020595555504163105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,512,0.01753422286775377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,512,0.017278222574128043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,512,0.05264800124698215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,128,0.002495999965402815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,12288,1.8453422122531469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,3072,0.009334222310119206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,6144,0.03799466623200311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,2048,0.016295111841625638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,16384,0.024314666787783306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,512,0.013819555441538492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,768,0.006239111224810283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,8192,0.026957333087921143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,3584,0.02574666672282749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,3584,0.02292355563905504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,3584,0.013434666726324292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,512,0.006569777925809224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,3584,0.007249777515729268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,1024,0.01646311084429423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,3072,0.014133334159851074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,16384,0.23046665721469453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,3584,0.023352000448438857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,6144,0.016178665889634024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,1536,0.016159999701711867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,512,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,512,0.027142221728960674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,65536,0.07226133346557617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,10240,0.0695635543929206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,1024,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,256,0.01609244445959727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,5120,0.22313332557678223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,256,0.030468444029490154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,3584,0.00759733302725686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,4096,0.027678221464157104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,6144,0.046201778782738574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,768,0.017495110630989075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,256,0.002528000002106031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,10240,0.21327643924289277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,8192,0.011147555377748279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,64,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,8192,0.06449688805474176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,8192,0.02131822208563487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,768,0.007311999797821045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,32,0.0021235555824306277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,32,0.005879111174080107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,6144,0.12769067287445068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,128,0.015824889143308003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,16384,0.3561760054694281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,3584,0.023696889479955036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,12288,0.09147555298275417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,6144,0.011877333124478659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,1536,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,768,0.004882666799757215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,256,0.014463999205165438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,2560,0.015788444214397006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,10240,0.007895111209816402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,3072,0.02199555602338579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,64,0.006225777582989798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,2048,0.004598222259018156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,5120,0.027801778581407335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,65536,0.036080890231662326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,512,0.013648889131016202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,65536,0.06788088877995808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,16384,0.0889857808748881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,4096,0.17074399524264863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,16384,0.027448000179396734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,16384,0.056503110461764865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,1536,0.008344000412358178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,512,0.01479555500878228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,6144,0.010665777656767102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,1536,0.003540444291300244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,65536,0.06494577725728352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,2048,0.008999110923873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,64,0.008965333302815756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,2048,0.0087333329849773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,256,0.016534222496880423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,5120,0.024308444725142583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,65536,0.07286666499243842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,3072,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,128,0.013820444544156393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,64,0.002171555534005165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,16384,0.03903822104136149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,512,0.012368000215954252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,768,0.014067555467287699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,768,0.017831110292010836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,10240,0.021934222843911912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,8192,0.03447377681732178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,10240,0.07906666729185316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,16384,0.06232444445292155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,6144,0.013425778183672162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,2048,0.004536889079544279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,1536,0.12020177311367458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,5120,0.009294222626421187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,4096,0.009500444763236577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,512,0.004917333523432414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,7168,0.026013332936498854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,16384,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,65536,0.006940444310506185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,16384,0.029667556285858154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,12288,0.024060444699393377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,3072,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,128,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,16384,0.04762044548988342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,16384,0.012048000262843238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,5120,0.027335110637876723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,5120,0.015111999379263984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,4096,0.05064177844259474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,1536,0.01517688896920946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,512,0.013064888616402945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,512,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,16384,1.9384977552625868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,2048,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,12288,0.019855111837387085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,65536,0.7062337663438586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,128,0.002130666747689247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,3584,0.009558222360081142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,12288,0.045026666588253446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,8192,0.16528533564673528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,16384,0.06839822398291694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,12288,0.29556889004177517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,256,0.009000000026490953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,6144,0.32359200053744847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,8192,0.06846400101979573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,1536,0.024127999941507976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,65536,0.6170719994439019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,12288,0.19772622320387098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,8192,0.024711999628278945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,768,0.014133334159851074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,12288,0.049061334795422025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,2048,0.01791822248035007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,2560,0.013427555561065674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,5120,0.027056889401541814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,1536,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,2048,0.02446488870514764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,512,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,64,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,1024,0.013215999636385175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,65536,0.1338497797648112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,1536,0.010781333678298526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,12288,0.01960711181163788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,2560,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,6144,0.021018667353524104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,16384,0.028259555498758953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,65536,0.09315644370185004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,3072,0.00637511122557852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,256,0.002562666725781229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,65536,0.15479289160834417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,5120,0.009444444543785518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,16384,0.08442311154471503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,12288,0.00868800034125646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,3072,0.007374222079912822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,65536,0.1240488953060574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,65536,0.0889822244644165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,10240,0.012454222473833295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,1536,0.016175111134847004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,768,0.01926399932967292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,5120,0.012492444780137805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,768,0.016826666063732572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,10240,0.017497777938842773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,7168,0.014430221584108142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,768,0.04420533445146349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,3072,0.013192888763215808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,512,0.014843554960356818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,256,0.005968889014588461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,128,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,1024,0.005604444278611078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,12288,0.01739377776781718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,1024,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,12288,0.029285334878497656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,4096,0.007732444339328342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,16384,0.006295111030340195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,4096,0.03151999910672506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,65536,0.07049600283304851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,5120,0.023061333431137934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,256,0.010495111346244812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,3072,0.023997333314683702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,768,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,10240,0.021340444684028625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,128,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,2560,0.006637333167923822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,32,0.001981333312061098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,768,0.03285422258906894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,8192,0.01279022213485506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,5120,0.5001164542304145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,8192,0.03543555405404832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,3584,0.007537778053018782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,128,0.0057964444988303715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,3584,0.01479644411140018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,64,0.0036640001667870414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,3072,0.02233955595228407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,256,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,32,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,6144,0.009010666774378883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,6144,0.01371022231049008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,768,0.01613333324591319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,3584,0.007647111184067196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,1536,0.004570666700601578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,2048,0.024951110283533733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,16384,0.02366755571630266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,3072,0.23130755954318574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,5120,0.18404000335269502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,3072,0.008628444539176093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,32,0.05205688873926798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,12288,0.048156443569395274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,128,0.0021751111166344746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,4096,0.017973333597183228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,512,0.005580444302823808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,65536,0.0588266650835673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,512,0.02537511123551263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,32,0.00254844439526399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,2560,0.01447111037042406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,1536,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,1024,0.015481778317027621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,10240,0.029532445801628962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,512,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,65536,0.3251715501149495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,6144,0.12481955687204997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,16384,0.015112888481881885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,16384,0.0407608879937066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,4096,0.009672000176376766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,7168,0.18716888957553438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,4096,0.010381333529949188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,2048,0.0058355554938316345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,5120,0.012488889197508493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,3072,0.01721244388156467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,256,0.014392889208263822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,8192,0.03858310977617899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,12288,0.030724445978800457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,7168,0.011377777490350934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,16384,0.2269617716471354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,8192,0.013248888982666863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,5120,0.0174666659699546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,256,0.023288889063729182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,256,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,32,0.0021191111041439907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,12288,0.16029422812991673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,256,0.012416000167528788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,6144,0.016968000266287062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,2560,0.009062222308582729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,32,0.0018613334331247543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,2560,0.007282666862010956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,5120,0.01219555570019616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,768,0.017502221796247695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,512,0.01035644445154402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,1536,0.010400888820489248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,2048,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,8192,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,12288,0.03065599997838338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,8192,0.027670222851965163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,12288,0.018604444132910836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,12288,0.11587377389272054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,256,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,256,0.014725334114498563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,6144,0.02184088859293196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,256,0.0038497779104444715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,5120,0.009354666703277165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,16384,0.024215110474162634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,65536,0.08865600162082249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,5120,0.021319111188252766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,2048,0.0909422238667806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,512,0.002834666727317704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,512,0.00527022240890397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,768,0.014832889040311178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,1536,0.0365608897474077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,6144,0.017902221944597032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,5120,0.008298666940795051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,12288,0.02126311096880171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,256,0.008667555948098501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,8192,0.05086844497256809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,16384,0.04778222242991129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,128,0.005030222237110138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,8192,0.01110400011142095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,2560,0.03223199976815118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,128,0.003863111138343811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,16384,0.20592266983456084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,256,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,2048,0.011191999746693505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,3584,0.04978933268123203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,1536,0.01514844430817498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,8192,0.012148444851239523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,10240,0.011769777370823754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,12288,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,1024,0.02463022205564711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,16384,0.03501333461867438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,128,0.016884444488419425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,7168,0.02335022224320306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,3584,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,128,0.010060444474220276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,4096,0.012519111235936483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,256,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,7168,0.018260444204012554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,1536,0.016864000095261466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,3584,0.09131377935409546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,16384,0.13060532675849065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,128,0.015824889143308003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,768,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,8192,0.021849777963426378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,8192,0.3987155490451389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,4096,0.03029777937465244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,8192,0.01037688884470198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,10240,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,3072,0.025278222229745653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,3584,0.01034755590889189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,768,0.015799999237060547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,10240,0.22633067766825357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,10240,0.038506666819254555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,6144,0.0848239991399977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,1536,0.00350222239891688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,10240,0.11354755030737983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,3072,0.0075351107451650835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,65536,0.07905955447090997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,16384,0.034033778640959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,10240,0.03565866748491923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,7168,0.1577848858303494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,8192,0.04149688945876227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,2560,0.00453333349691497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,12288,0.2878986729515923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,1024,0.013761777844693927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,65536,0.9359075758192273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,65536,0.1281813383102417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,5120,0.01719288859102461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,2560,0.005203555441564984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,10240,0.016743110285864938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,128,0.024919110867712233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,3584,0.019959999455346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,12288,0.03562488820817735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,5120,0.02237066626548767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,4096,0.08289600080913968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,16384,0.041519999504089355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,3072,0.07690044244130452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,128,0.007241778075695038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,3584,0.011563555234008364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,1024,0.007891555627187094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,256,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,256,0.018596443865034316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,128,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,512,0.0025902222841978073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,1024,0.0028400001012616684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,65536,0.36844711833530003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,1536,0.01738755570517646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,1024,0.007359999749395583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,8192,0.015537778536478678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,256,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,1024,0.007333333293596904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,6144,0.00943199959066179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,4096,0.007610666255156199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,256,0.002507555608948072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,8192,0.02148888839615716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,1536,0.006535111202134027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,7168,0.01924533314175076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,6144,0.06363466713163587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,4096,0.05224177903599209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,16384,0.05396622088220385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,768,0.06032622522777981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,4096,0.008054222497675154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,1024,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,64,0.002393777792652448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,3072,0.006913777854707506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,12288,0.05941066477033827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,1024,0.013739555246300168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,10240,0.043975111511018544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,3584,0.008278222547637092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,8192,0.01807733376820882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,2048,0.05358577768007914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,6144,0.03355822298261855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,10240,0.01792266633775499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,2048,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,768,0.010978666444619497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,512,0.004943999979231092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,1536,0.015456888410780164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,4096,0.00452977791428566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,512,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,4096,0.024678222007221643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,2560,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,10240,0.03152622116936578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,1024,0.06185422341028849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,256,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,512,0.009642666412724389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,512,0.01241422196229299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,8192,0.02773333258099026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,10240,0.4342906739976671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,10240,0.06915466652976142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,64,0.0075733330514695905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,32,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,4096,0.004552000098758274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,16384,0.03151199883884854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,10240,0.015453333655993143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,4096,0.007605333295133378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,256,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,5120,0.01127822200457255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,16384,0.030814223819308813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,2048,0.004327110946178436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,4096,0.017507554756270517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,256,0.009333333207501305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,128,0.012720000412729053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,4096,0.16824532879723442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,4096,0.06061155266231961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,256,0.014828445182906257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,1536,0.012105777859687805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,32,0.006311111152172089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,128,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,3584,0.0075804442167282104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,1536,0.01202400028705597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,1536,0.006008000009589725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,2560,0.012111110819710625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,2560,0.025411556164423626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,512,0.002768888862596618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,3584,0.010913777682516309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,16384,0.12650222248501247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,512,0.0151582227812873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,512,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,10240,0.004545777622196409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,10240,0.9751422670152453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,65536,0.4980826907687717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,5120,0.009317333499590555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,10240,0.3382524384392633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,128,0.012243555651770698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,12288,0.016176000237464905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,1024,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,2048,0.0151235560576121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,512,0.010681777364677854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,16384,0.010969777901967367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,6144,0.0110880004035102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,768,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,4096,0.06256444586647882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,12288,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,12288,0.012764444781674279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,64,0.0032248888164758682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,512,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,10240,0.03853333327505324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,1536,0.005967111223273807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,3072,0.09050133493211533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,4096,0.011887110769748688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,10240,0.02533066603872511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,1024,0.014510222607188754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,2560,0.07502755853864881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,5120,0.026064000195927087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,2560,0.07748711109161377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,8192,0.061547557512919106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,32,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,12288,0.03772266705830892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,768,0.005883555445406172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,2048,0.018179555733998615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,1024,0.007596444752481248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,5120,0.02536711096763611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,2048,0.012072000238630505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,1024,0.00656444455186526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,7168,0.03190844588809543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,5120,0.01129955550034841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,3072,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,32,0.002475555572244856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,16384,0.04406933320893181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,7168,0.019333332777023315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,2048,0.013803555733627744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,6144,0.009014222357008193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,3584,0.007981333467695449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,8192,0.01833866702185737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,6144,0.02199911077817281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,12288,0.04177600145339966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,8192,0.014679110712475248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,5120,0.008662222160233391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,2560,0.004872888740566042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,3584,0.010694222317801582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,3072,0.005304000029961268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,512,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,768,0.008347555167145198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,8192,0.046762665112813316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,1536,0.06883733140097724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,16384,0.04400444361898634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,1024,0.014097778333557976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,256,0.01646133263905843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,6144,0.006235555642180973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,6144,0.010381333529949188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,1536,0.11176621913909912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,8192,0.006257777826653586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,65536,0.09292622407277425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,8192,0.038513776328828596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,1024,0.0075342224703894714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,10240,0.027079110344250996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,1536,0.02381155557102627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,32,0.011357333097192975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,256,0.006572444405820634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,64,0.002128888956374592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,8192,0.018608889646000333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,3072,0.016591999265882704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,3072,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,2560,0.012684444586435953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,8192,0.01035377797153261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,2560,0.007330666813585494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,2560,0.2376035584343804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,2560,0.03323466579119364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,3072,0.022340445054901972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,2048,0.00796088907453749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,3584,0.184207108285692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,128,0.0024968888610601425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,12288,0.060749332110087075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,3072,0.024420445164044697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,512,0.004898666507667965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,8192,0.05836533175574409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,5120,0.017143999536832173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,1024,0.005636444522274866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,6144,0.016081778539551627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,5120,0.05073777834574381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,2560,0.016187555260128446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,16384,0.06910933388604058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,2048,0.012362666428089142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,64,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,1536,0.013755555782053204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,6144,0.01888799998495314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,64,0.001839111041691568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,256,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,65536,0.04733333322736952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,1536,0.01621955633163452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,6144,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,3072,0.02882044514020284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,12288,0.04968800147374471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,256,0.04145244426197476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,8192,0.023013333479563396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,5120,0.038374221987194486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,1536,0.014801777071423002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,2560,0.0070186663005087115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,2560,0.022152000003390845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,16384,0.8704409069485135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,6144,0.014751111467679342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,6144,0.025519111090236243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,2048,0.006603555546866522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,32,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,64,0.005682666682534748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,64,0.002499555548032125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,8192,0.026307556364271376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,3072,0.026075555218590632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,64,0.006611555400821898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,8192,0.01647911138004727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,4096,0.07819466458426581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,12288,0.0590462221039666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,12288,0.017488888568348356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,10240,0.020577778418858845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,256,0.014781332678265043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,12288,0.11108089155620998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,1024,0.005942222144868638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,3584,0.09999022218916152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,6144,0.021432888176706102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,4096,0.0651786658498976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,7168,0.012993777791659037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,65536,0.07263911432690091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,5120,0.0358897778722975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,3072,0.008961777720186446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,768,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,512,0.002826666666401757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,7168,0.02770311137040456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,256,0.016851555969980028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,768,0.0028515555378463534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,10240,0.018216888109842937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,3584,0.01891377733813392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,768,0.005655999812814925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,2048,0.07717866367763944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,1536,0.03343733482890659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,65536,0.08624888790978326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,6144,0.024881778491867915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,3072,0.0075866662793689305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,1536,0.006162666612201267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,16384,0.01924800044960446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,12288,0.026604443788528442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,6144,0.004862222406599256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,3584,0.007546666595670912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,2560,0.006858666737874349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,4096,0.0042462220622433555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,6144,0.13875733481513128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,12288,0.06646755668852064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,3072,0.012069332930776807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,1536,0.006573333508438534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,3584,0.01517688896920946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,768,0.0032844444115956626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,16384,0.03719377848837111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,65536,0.11904177400800918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,8192,0.016221332881185744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,256,0.006263999889294307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,3584,0.018372444642914664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,512,0.010029333333174387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,128,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,256,0.002495999965402815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,64,0.007273777491516537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,768,0.013426666458447775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,512,0.005208000126812193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,16384,0.015415110521846347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,128,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,1024,0.016005333926942613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,5120,0.010692444112565784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,2560,0.006963555183675554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,256,0.02123111155298021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,32,0.002185777864522404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,32,0.01000711073478063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,2560,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,768,0.005684444473849402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,10240,0.14001422458224827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,768,0.016191111670600045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,32,0.009668444593747457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,4096,0.18447911739349365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,6144,0.3608240021599664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,1536,0.012890666723251343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,1024,0.017885333961910672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,7168,0.014828445182906257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,2560,0.012116444607575735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,8192,0.031939556201299034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,12288,0.030975110001034204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,3072,0.009682666924264696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,64,0.002185777864522404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,7168,0.012726222475369772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,2048,0.0515288876162635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,256,0.009301332963837517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,4096,0.02774311105410258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,3072,0.01001422190003925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,3584,0.01650399963061015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,128,0.005888000130653381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,256,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,512,0.0024968888610601425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,65536,0.08244088623258802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,512,0.011417778001891242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,512,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,6144,0.023709333605236475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,128,0.0038702223036024305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,7168,0.013802666631009845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,1024,0.021310221817758348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,768,0.00319377767542998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,2560,0.008279110822412703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,256,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,10240,0.12881955835554335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,2560,0.01421866648726993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,4096,0.013105777402718862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,10240,0.012110222544935016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,768,0.007093333535724216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,64,0.015806222955385845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,2560,0.006584888945023219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,128,0.013376000026861826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,2560,0.015114666687117683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,768,0.005948444621430502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,10240,0.036871112055248685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,2560,0.015390222271283468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,6144,0.013474666410022311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,4096,0.025516443782382544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,768,0.022078222698635522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,10240,0.0361057784822252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,7168,0.01811555524667104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,2560,0.0074275558193524676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,512,0.00720888872941335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,10240,0.5703040228949653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,7168,0.02238133384121789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,16384,0.0325271122985416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,8192,0.01886755559179518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,2560,0.02104266650146908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,10240,0.010575999816258749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,512,0.0025048889219760895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,4096,0.02808088726467556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,12288,0.01648622254530589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,256,0.002477333363559511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,5120,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,1536,0.006186666587988536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,64,0.0018177777528762817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,256,0.016501333978441026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,1024,0.006940444310506185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,5120,0.00865955568022198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,512,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,3072,0.007281777759393056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,3072,0.009553777674833933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,1024,0.006913777854707506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,3072,0.004927111168702443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,128,0.014314666390419006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,65536,0.47331375545925564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,5120,0.012158222496509552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,12288,0.19701066282060412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,16384,0.015474667151769003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,256,0.006248000181383557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,12288,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,65536,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,4096,0.016768000192112394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,3072,0.007120889094140794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,512,0.011383111278216044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,64,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,2560,0.05729155408011543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,10240,0.011716444459226398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,16384,0.015470221638679504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,6144,0.015174221661355762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,128,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,3072,0.005894222193294101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,256,0.004960888789759742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,12288,0.07431644201278687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,512,0.015829333000712924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,4096,0.004571555389298333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,8192,0.07804622252782185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,1536,0.008303110798199972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,128,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,3584,0.007640889121426477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,1024,0.0035093331502543557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,12288,0.029346668057971533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,3584,0.008932444784376357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,7168,0.0613973339398702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,128,0.014779556128713818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,3072,0.007015110717879401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,4096,0.012327111429638334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,6144,0.07874133189519246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,8192,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,12288,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,1024,0.021320000290870667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,5120,0.008634666601816813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,2560,0.004926222066084544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,256,0.008169777691364288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,3072,0.019308444526460435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,7168,0.04411733481619093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,256,0.015450666348139444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,6144,0.01350488927629259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,3584,0.007635555333561367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,128,0.005946666830115848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,1536,0.008192000289758047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,3072,0.09891821940739949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,12288,0.04134666588571336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,1024,0.012406222522258759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,512,0.031345778041415744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,3072,0.01350311107105679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,256,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,1024,0.01093066649304496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,65536,0.057670222388373486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,10240,0.03268533282809787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,3072,0.0420684450202518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,3584,0.0162444445821974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,12288,0.02658222284581926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,2048,0.006629333313968446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,4096,0.00424799985355801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,512,0.033118221494886614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,4096,0.007920000288221572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,3584,0.02607911162906223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,2560,0.004853333450025982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,1536,0.008079110748238033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,256,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,7168,0.01163111130396525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,768,0.024340444140964087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,256,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,3072,0.010781333678298526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,6144,0.1242248879538642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,256,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,1536,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,128,0.0024622221373849446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,4096,0.004287999951177173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,5120,0.004888888862397936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,7168,0.01310222182008955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,7168,0.014108444253603617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,65536,0.3457635508643256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,32,0.017543110582563613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,16384,0.057892441749572754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,32,0.004701333327425851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,6144,0.05249866512086657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,1536,0.006311999840868845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,1536,0.013640888863139682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,4096,0.008355555435021719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,16384,0.022346667117542688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,7168,0.047862221797307335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,2560,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,10240,0.03721066647105747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,2048,0.01817244456874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,2048,0.007304888632562425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,5120,0.008550222549173567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,4096,0.006598222172922558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,128,0.0021742222209771476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,65536,0.07974577612347074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,2560,0.00527022240890397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,12288,0.00869511150651508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,3072,0.16275021764967176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,768,0.0030746666921509635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,2560,0.02983377708329095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,12288,0.02810488806830512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,3072,0.009912889036867354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,10240,0.01812799937195248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,2048,0.0584755539894104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,2560,0.022597332795461018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,2048,0.03458044595188565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,128,0.013776000175211163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,4096,0.02879377868440416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,4096,0.02188088827662998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,8192,0.00590133335855272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,128,0.019896888070636325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,8192,0.16065511438581678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,2048,0.016900445024172466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,2560,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,4096,0.017232000827789307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,3584,0.007725333174069722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,8192,0.02365511159102122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,2048,0.006630222416586346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,5120,0.10199111037784153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,5120,0.014557333456145393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,10240,0.4324977662828233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,2048,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,4096,0.1740044487847222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,7168,0.013707555830478668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,128,0.00216977774269051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,1536,0.015531554818153381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,8192,0.6110684606764052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,4096,0.08226755592558119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,16384,0.02980177932315403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,1024,0.020281778441535104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,7168,0.044569777117835156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,7168,0.105277332994673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,7168,0.010066666536860997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,3584,0.08823021915223862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,8192,0.2712328963809543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,3072,0.021736888421906367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,512,0.005582222094138463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,768,0.01829866733815935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,1536,0.014467555615637036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,3584,0.1654355525970459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,16384,0.15294488271077475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,3584,0.012401777837011548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,6144,0.009329777624871995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,128,0.005521777603361342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,65536,0.09405066569646199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,768,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,7168,0.011028444601429833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,256,0.004914666629499859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,7168,0.011404444773991903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,2048,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,512,0.0052844443255000645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,12288,0.03592977921168009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,256,0.015828443898095023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,128,0.00508266646001074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,768,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,10240,0.16561599572499594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,768,0.004556444370084339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,2560,0.004578666554556953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,6144,0.011707555916574268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,12288,0.03421955638461643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,3584,0.009088888764381409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,16384,0.5491440031263564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,6144,0.062374220954047314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,8192,0.0302471121152242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,5120,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,768,0.011060444845093621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,8192,1.3165253533257377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,256,0.006636444479227066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,1024,0.007999999655617608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,10240,0.014988443917698331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,32,0.005912888795137405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,65536,0.010667555862002902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,2048,0.014456000592973499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,2560,0.007307555940416124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,1024,0.004551110996140374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,6144,0.004922666483455234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,10240,0.020580444071027968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,512,0.005424889011515512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,2048,0.005440000030729506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,1024,0.02532355487346649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,1536,0.007655999726719326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,16384,0.045864890019098915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,2048,0.006687110910813014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,768,0.03626311156484816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,1536,0.0053182223604785065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,3584,0.011172444456153445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,3072,0.0105262224872907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,65536,0.08281955454084608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,768,0.005333333379692501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,10240,0.012468444804350534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,256,0.011767999993430244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,2048,0.01643200053109063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,3584,0.014115555418862237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,768,0.013767110804716745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,1024,0.011377777490350934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,5120,0.008623999853928884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,5120,0.00959288908375634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,256,0.002473777780930201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,2048,0.004261333495378494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,12288,0.017120000388887193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,12288,0.016754666964213055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,3584,0.015125332607163323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,768,0.0028497777465316984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,2048,0.05102755626042684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,65536,0.01995022263791826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,65536,0.24030399322509766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,65536,0.05542844533920288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,3584,0.0274311105410258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,8192,0.01886577738655938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,1024,0.005864000154866113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,768,0.005544888890451855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,65536,0.1037822233306037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,32,0.0018719999740521114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,2048,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,7168,0.013403555585278405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,2560,0.006853333363930385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,6144,0.009286222358544668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,1536,0.006255110932721033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,3072,0.009349333743254343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,4096,0.05888089206483629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,64,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,2560,0.039334221018685236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,12288,0.012715555727481842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,12288,0.04260444309976366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,2560,0.005693333430422678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,8192,0.02236355510022905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,1024,0.00739911115831799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,2048,0.004513777792453766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,512,0.005197333378924264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,1024,0.008512000242869059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,1536,0.01754755609565311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,32,0.0021297778520319196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,2560,0.013160000244776407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,128,0.01295111080010732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,3072,0.028453333510292902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,1024,0.00600888869828648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,7168,0.04593866732385424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,2560,0.00693866651919153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,3584,0.010694222317801582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,1536,0.00351200004418691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,65536,0.05819111400180393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,128,0.002483555633160803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,256,0.00490133340160052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,3584,0.012442666623327466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,16384,0.09988444381290013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,2048,0.006650666809744305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,4096,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,3072,0.008660444782839881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,768,0.004935111022657818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,2048,0.004229333251714706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,128,0.011390222443474663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,65536,0.09101600117153591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,256,0.005937777873542573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,16384,0.4071253405676947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,256,0.023943111300468445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,10240,0.005714666512277391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,512,0.002530666689078013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,12288,0.03403110967742072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,10240,0.03142933381928338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,5120,0.0086666668454806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,16384,0.2767164442274305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,12288,0.03468177715937296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,10240,0.312300443649292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,7168,0.04837866624196371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,1024,0.038380444049835205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,3584,0.04034311241573758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,65536,0.111117336485121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,10240,0.01514311134815216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,1536,0.01037333326207267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,2560,0.01606666710641649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,128,0.03638311227162679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,8192,0.04839644498295254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,3584,0.030849777989917334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,5120,0.008423111504978603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,64,0.0032844444115956626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,256,0.006096000058783426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,6144,0.033575998412238225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,7168,0.14768711725870767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,512,0.00591822216908137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,7168,0.01089599976936976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,256,0.007712889048788283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,1536,0.024302222662501868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,256,0.002530666689078013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,8192,0.03915199968549941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,2560,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,12288,0.05382133192486233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,2560,0.11320800251430935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,1536,0.018598222070270114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,3584,0.00675999994079272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,1536,0.008024888734022776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,1024,0.0029440000653266907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,256,0.0021724444296624926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,3584,0.013794666363133324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,64,0.007682667010360294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,1536,0.00600622221827507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,12288,0.20161688327789307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,3072,0.08718399869071113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,6144,0.009709333380063375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,512,0.005962666538026597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,5120,0.010135999984211391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,2560,0.006671111202902264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,32,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,65536,0.12149511443244086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,2048,0.012453333371215396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,12288,0.27101156446668834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,128,0.0032808888289663526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,4096,0.01423822177780999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,3072,0.11105333434210883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,1024,0.01310755560795466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,256,0.007823111282454597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,512,0.007998222278224098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,2048,0.014427555931939019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,2048,0.007304000357786815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,12288,0.02407377792729272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,3072,0.02186311119132572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,64,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,4096,0.007319110963079665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,65536,0.12219911151462132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,1536,0.05885244740380181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,2560,0.009207111265924241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,7168,0.019647111495335896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,64,0.010583110981517367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,5120,0.009013333254390294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,1024,0.015784000356992085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,4096,0.020630222227838304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,512,0.006908444480763541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,8192,0.015951111912727356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,5120,0.010620444185203977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,3584,0.008271999657154083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,4096,0.02201866606871287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,6144,0.004882666799757215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,3584,0.03534311056137085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,7168,0.023342221975326538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,2560,0.006907555378145642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,2560,0.007296889192528195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,12288,0.20755467149946424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,768,0.01888799998495314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,1024,0.015844444433848064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,12288,0.017255110873116385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,2560,0.023425777753194172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,12288,0.022297778063350256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,256,0.00282577777074443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,4096,0.014195554786258273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,10240,0.06259377797444661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,3072,0.035736001200146146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,1536,0.006234666539563074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,2560,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,512,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,3072,0.015447111593352424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,2048,0.028804444604449805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,64,0.001850666685236825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,10240,0.2110897832446628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,10240,0.03287200133005778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,128,0.016128889388508268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,4096,0.013560000393125745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,1536,0.01590577761332194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,3072,0.007979555262459649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,1536,0.01886044442653656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,6144,0.44157155354817706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,3072,0.10178755389319526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,65536,0.06823911269505818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,8192,0.01070755554570092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,2048,0.011095110740926532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,6144,0.027293332748942908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,768,0.008998221821255153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,2048,0.10261778036753337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,65536,0.09219111336602105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,768,0.017218665944205392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,32,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,8192,0.015274667077594332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,3072,0.008611555728647444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,5120,0.026138666603300307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,512,0.02459555533197191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,2048,1.2314382129245334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,65536,0.07039199935065375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,2048,0.018595554762416415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,65536,0.2132880025439792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,6144,0.06652888986799452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,10240,0.03520888752407498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,7168,0.015471110741297403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,512,0.0028373334142896864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,1024,0.06659111049440172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,2048,0.015804444750150044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,5120,0.017843556072976854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,2048,0.004505777938498391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,2048,0.014167999227841696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,1536,0.00590844452381134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,1024,0.012418666647540199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,768,0.004625777817434735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,32,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,12288,0.044663998815748424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,2560,0.02611733310752445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,65536,0.10801955726411606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,7168,0.01651022169325087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,6144,0.02574666672282749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,10240,0.058584888776143394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,12288,0.00997688869635264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,65536,0.10085688696967231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,8192,0.04551733202404446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,2560,0.004879999905824661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,32,0.0021199999998013177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,2560,0.012808889150619507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,12288,0.017839110559887357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,4096,0.02106844385464986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,64,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,3072,0.4324728912777371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,4096,0.010755555497275459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,16384,0.03021866745418972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,6144,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,768,0.0032133333798911837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,8192,0.03537866804334853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,1024,0.006232888748248418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,65536,0.054543998506334096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,2560,0.020971554848882887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,1536,0.014821334017647637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,2560,0.012822222378518848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,3584,0.02515022291077508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,5120,0.008653333617581261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,512,0.00519466648499171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,12288,0.014141332772043018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,512,0.0042160000238153665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,8192,0.017342221405771043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,128,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,64,0.016933333542611863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,64,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,6144,0.009823111196359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,10240,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,5120,0.007974222302436829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,5120,0.008664000365469191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,3584,0.012039110892348819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,4096,0.021307556165589228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,6144,0.013760888742076026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,4096,0.011360000405046674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,2560,0.006973333656787872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,1536,0.004573333180612988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,512,0.002492444382773505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,768,0.02498666610982683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,4096,0.046931554873784385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,256,0.005622222191757626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,1024,0.007002666592597961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,64,0.005879999862776862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,6144,0.030395554171668157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,256,0.014482667048772177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,3584,0.022965333527988855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,512,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,6144,0.011399110986126794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,16384,0.023079999619060095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,4096,0.03931466738382975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,10240,0.03108888864517212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,32,0.012076444096035428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,32,0.0019662222928471034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,16384,0.04264000058174133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,16384,0.08409600125418769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,7168,0.01071111112833023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,32,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,768,0.0029164445069101122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,65536,0.0833608905474345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,64,0.0019004444281260173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,2560,0.006665777828958299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,2048,0.012455110748608908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,12288,0.027823110421498615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,3584,0.2984284559885661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,16384,0.049676444795396596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,128,0.013768889009952545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,6144,0.11716622776455349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,2560,0.01958755486541324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,6144,0.028796444336573284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,8192,0.15837155448065862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,2048,0.05378133389684889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,768,0.010999999940395355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,16384,0.05770044194327461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,1536,0.003495111233658261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,2048,0.00793688909875022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,768,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,3072,0.008740444150235917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,12288,0.1835235489739312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,7168,0.03838488790724013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,16384,0.03700444433424208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,3584,0.007272889216740926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,3584,0.0148053334818946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,4096,0.018216888109842937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,3584,0.007616000043021307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,7168,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,2048,0.022847111026446026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,1024,0.022143999735514324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,1024,0.0041511112617121804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,64,0.006236444330877728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,10240,0.020276443825827706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,5120,0.027744000156720478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,5120,0.029319110843870375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,1536,0.00591111100382275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,3072,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,256,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,256,0.011390222443474663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,5120,0.010271111296282874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,2048,0.009927999642160203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,128,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,256,0.007982222570313348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,7168,0.011383111278216044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,4096,0.186408003171285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,32,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,128,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,3584,0.007999999655617608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,65536,0.164375106493632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,16384,0.04646133383115133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,512,0.021743110484547083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,7168,0.013056889176368713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,128,0.002165333264403873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,64,0.006187555690606435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,16384,0.035635554128222995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,128,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,12288,0.13026400407155356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,1536,0.011691555380821228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,65536,1.8483492533365886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,6144,0.01605333387851715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,256,0.005054222212897407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,10240,0.21894578138987222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,6144,0.009822222093741098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,4096,0.007618666523032718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,3584,0.01588355501492818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,256,0.005990222096443176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,1536,0.012077333198653327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,2560,0.012392000191741519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,16384,0.0192577772670322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,256,0.006629333313968446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,512,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,128,0.10066133075290257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,16384,0.027582221561008032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,512,0.012441777520709567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,2560,0.015126221709781222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,5120,0.012864000267452665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,1024,0.0031475555151700974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,512,0.002524444419476721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,3584,0.007669332954618666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,8192,0.01756000022093455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,768,0.00963199966483646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,4096,0.01516888870133294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,10240,0.032758222685919866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,3072,0.021707556313938562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,512,0.012865777644846173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,64,0.052001777622434825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,8192,0.03864266806178623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,6144,0.01772444446881612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,7168,0.015040889382362366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,10240,0.019229332605997723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,2048,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,7168,0.21681777636210123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,32,0.01238933371173011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,6144,0.10119111008114284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,512,0.005758222192525864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,5120,0.05404799845483568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,32,0.007350222104125553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,8192,0.06665066878000896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,16384,0.28448979059855145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,8192,0.04223733478122287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,2560,0.007250666618347168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,3584,0.018239110708236694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,3072,0.0459191103776296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,65536,0.03947822252909342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,2560,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,6144,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,2560,0.00721688899728987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,768,0.01679999960793389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,3584,0.004870222260554631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,1024,0.006219555520349079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,64,0.0018613334331247543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,8192,0.3407786687215169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,6144,0.20750666989220512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,12288,0.017083555459976196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,65536,0.07650399870342679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,768,0.015457777513398064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,2560,0.007168889045715332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,12288,0.044080889887279935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,256,0.016171556380059984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,512,0.013795555465751223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,2048,0.06567111280229357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,1024,0.02201333310869005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,1024,0.04062844316164652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,8192,0.09915377696355183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,32,0.01550044450494978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,64,0.0018657777044508194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,512,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,1536,0.003181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,512,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,12288,0.01812355551454756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,256,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,3584,0.010844444235165914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,3584,0.01793066660563151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,5120,0.11771999465094672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,4096,0.007619555625650618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,768,0.002479111154874166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,128,0.002834666727317704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,5120,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,5120,0.009309333231714036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,8192,0.010228444304731157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,16384,0.010675555302037133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,1024,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,3072,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,65536,0.15862133767869738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,5120,0.03320088982582092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,64,0.005195555587609609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,1024,0.006232000059551663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,65536,0.10740888781017727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,5120,0.060608002874586314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,3584,0.007323555648326874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,12288,0.01812622282240126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,3584,0.021759111020300124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,1024,0.005763555566469829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,128,0.005197333378924264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,6144,0.01683111157682207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,6144,0.034500443273120456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,64,0.0018097776919603348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,6144,0.009316444396972656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,2560,0.05061422122849358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,4096,0.013739555246300168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,2048,0.006242666807439592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,1536,0.00500266667869356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,3584,0.007928888830873702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,65536,0.07649955484602186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,16384,0.0248391115003162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,10240,0.10240533616807725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,32,0.00812355594502555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,65536,0.1467404497994317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,512,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,4096,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,768,0.01260444439119763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,256,0.06620089213053386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,6144,0.009347555538018545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,768,0.012711111042234631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,7168,0.014583110809326172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,16384,0.042444444364971586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,1024,0.003569777641031477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,8192,0.010492444038391113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,4096,0.01483022173245748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,64,0.040298667218950056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,2560,0.006631111105283101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,3072,0.014801777071423002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,2560,0.0748622218767802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,1536,0.1034266683790419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,7168,0.022590221630202398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,1024,0.00756533361143536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,768,0.025640888346566096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,768,0.006800000038411882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,256,0.004951111144489712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,128,0.015229332778188916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,65536,0.1497288942337036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,6144,0.0059573331640826324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,2048,0.004591111093759537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,2048,0.016172443826993305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,10240,0.012732444538010491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,1024,0.005624889085690181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,768,0.00423200014564726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,2048,0.07624977827072144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,4096,0.009385777844323052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,16384,0.019645333290100098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,3584,0.007348444726732042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,512,0.015231110983424716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,8192,0.04324977927737766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,3072,0.010767999622556897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,12288,0.0429422226217058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,7168,0.0337137778600057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,65536,0.07598400115966797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,64,0.28567645284864635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,2048,0.006188444379303191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,3584,0.011596444580290051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,7168,0.16535645061069065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,128,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,1536,0.016535111599498324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,256,0.01223288890388277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,256,0.005349333501524395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,256,0.008594666918118795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,2560,0.012891555825869242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,3584,0.009993777506881291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,4096,0.044847998354170054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,128,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,10240,0.016812443733215332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,6144,0.009711999860074785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,65536,0.16055111090342203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,10240,0.015459555718633862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,256,0.016490666402710807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,6144,0.028736889362335205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,768,0.01733688844574822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,3584,0.01720177796151903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,1536,0.01962311069170634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,512,0.011621333658695221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,5120,0.01108977778090371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,1536,0.005892444401979446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,3584,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,16384,0.055996444490220815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,512,0.004578666554556953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,512,0.005892444401979446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,8192,0.010710222025712332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,2560,0.007171555525726742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,5120,0.006258666515350342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,512,0.013085333009560903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,1024,0.005508444375462002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,128,0.010383110907342698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,12288,0.01072266697883606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,65536,4.609535217285156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,1024,0.008021333151393467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,2560,0.0580257773399353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,3584,0.010997333460383944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,12288,0.037802666425704956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,1024,0.0403128895494673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,12288,0.005515555540720622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,32,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,64,0.0028159999185138275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,3072,0.0058400001790788435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,8192,0.012254222399658628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,4096,0.020310221446885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,1024,0.0032248888164758682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,6144,0.009300444689061906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,2560,0.006953777538405524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,512,0.012764444781674279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,12288,0.024705777565638225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,1536,0.012619555824332766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,2560,0.11632443798912896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,1536,0.006159111029571957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,1536,0.014707555373509726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,12288,0.022632888621754114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,5120,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,3584,0.04460533459981283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,10240,0.11112799909379746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,128,0.004898666507667965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,8192,0.016766221986876596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,65536,0.14652266767289904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,8192,0.024762666887707178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,1536,0.027479999595218237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,5120,0.0462337765428755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,6144,0.00693866651919153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,1536,0.014448000325096978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,7168,0.009043555292818282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,65536,0.10331644614537557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,128,0.004972444640265571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,5120,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,65536,0.0548622210820516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,65536,0.08981333176294963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,3072,0.01481244464715322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,3072,0.004634666774008009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,1536,0.019540444016456604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,5120,0.0094355551732911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,5120,0.010048000348938836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,16384,0.0769760012626648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,1536,0.004425777743260066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,3072,0.004576888763242298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,2560,0.008309333688682979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,3072,0.007363555332024892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,8192,0.022052443689770166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,256,0.01443822185198466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,512,0.017839999662505258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,1024,0.012225777738624148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,4096,0.017837334010336135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,65536,0.2135822243160672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,10240,0.045992887682384916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,4096,0.04120799899101257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,1536,0.024029332730505202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,5120,0.007598222129874759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,32,0.0031848889258172777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,2048,0.022876444790098403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,7168,0.15422756142086452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,512,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,8192,0.02033866610791948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,512,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,16384,0.06255822049246894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,1536,0.06936533583535089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,32,0.007660444411966536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,1536,0.00590133335855272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,4096,0.00795288880666097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,512,0.005184000151024925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,256,0.0024951110697454875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,8192,0.019268444842762418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,64,0.002528888897763358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,2560,0.01647555496957567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,4096,0.01588800052801768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,64,0.003588444242874781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,256,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,3584,0.01756888959142897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,512,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,2048,0.010319111247857412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,1536,0.006242666807439592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,65536,0.13156533241271973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,2048,0.019727110862731934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,3584,0.007997333175606197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,256,0.0023786665664778813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,16384,0.04273955689536201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,3072,0.007631999750932057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,2560,0.019336889187494915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,10240,1.5589715109931097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,32,0.005511111269394557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,65536,0.07999733421537611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,2048,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,512,0.0029333333174387612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,768,0.0028640000770489373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,64,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,64,0.002481777841846148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,16384,0.5068995687696669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,1536,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,16384,0.38449690077039933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,12288,0.14966844187842474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,1024,0.005888888819350137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,16384,0.05330311258633932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,1536,0.0069191112286514705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,8192,0.010676444404655032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,64,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,3072,0.00943377779589759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,6144,0.05022933416896396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,6144,0.01995733380317688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,1536,0.003556444413132138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,8192,0.01790488925245073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,2048,0.009280000295903947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,64,0.0022124445272816554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,2048,0.016560888952679105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,512,0.005888000130653381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,4096,0.007672000262472365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,128,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,2048,0.005933333188295364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,3584,0.008610666626029545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,64,0.002040000011523565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,1024,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,64,0.007300444775157505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,3072,0.03813955518934462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,16384,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,65536,0.05514666769239637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,7168,0.00869422240389718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,10240,0.05137155453364054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,10240,0.10183822446399265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,256,0.016498666670587327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,128,0.0034862222770849862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,1024,0.0029093333416514923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,4096,0.008299555215570662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,8192,0.02945244312286377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,3584,0.018787556224399142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,1536,0.01870133313867781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,6144,0.004953777624501122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,8192,0.014538667268223233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,128,0.004928888960017098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,768,0.005583111196756363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,1536,0.014124444789356656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,128,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,12288,0.004886222382386525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,2048,0.020291555258962844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,2560,0.006627555522653792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,3072,0.02910400099224514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,256,0.002451555596457587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,3072,0.014536889062987434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,3072,0.009979555176364051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,7168,0.01036088913679123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,16384,0.03541422221395705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,2048,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,256,0.0028328889360030494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,5120,0.016318221886952717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,4096,0.03102933367093404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,2048,0.0042035554846127825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,10240,0.016985777351591323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,3072,0.08288000027338664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,6144,0.011577777564525604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,7168,0.021254221598307293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,128,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,8192,0.020111110475328233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,6144,0.008966222405433655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,4096,0.006569777925809224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,6144,0.02200711104604933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,7168,0.022409778502252366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,512,0.011735999749766456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,8192,0.18349777327643502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,10240,0.0162435554795795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,64,0.002473777780930201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,10240,0.048807111051347524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,16384,0.2376453346676297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,1024,0.004271111140648524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,1536,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,6144,0.017858665850427415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,3072,0.006934222247865465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,64,0.005533333453867171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,6144,0.021371554997232225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,5120,0.012275555895434486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,12288,0.06648355722427368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,1024,0.023775999744733173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,5120,0.03941599859131707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,8192,0.0399911105632782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,16384,7.004364437527126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,65536,0.22055555714501274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,16384,0.11500622166527642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,3072,0.03491288754675124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,64,0.005924444645643234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,512,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,2560,0.007309333317809635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,7168,0.016083555089102853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,768,0.015398222539159985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,8192,0.02791822287771437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,64,0.0018648888087934917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,4096,0.021368889345063105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,8192,0.5613457891676161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,512,0.021280889709790547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,8192,0.04182755615976122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,128,0.001810666587617662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,16384,0.05568799707624647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,256,0.010667555862002902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,512,0.013756444056828817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,4096,0.010558221903112201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,16384,0.005624889085690181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,2048,0.006562666760550604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,8192,0.15323289235432944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,2048,0.006263999889294307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,10240,0.005948444621430502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,3584,0.007552000383536021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,512,0.01482222146458096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,65536,0.2939217885335286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,16384,0.04708533485730489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,3584,0.007266666326257918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,12288,0.013125333521101209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,16384,0.3180204497443305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,7168,0.011239111423492432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,5120,0.009362666971153682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,768,0.11253955629136826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,4096,0.014456000592973499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,12288,0.18889689445495605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,6144,0.02404088940885332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,8192,0.011719110939237805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,5120,0.008192888564533657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,8192,0.02125066684352027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,12288,0.018199111024538677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,768,0.009043555292818282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,10240,0.021336000826623704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,128,0.012402666939629449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,512,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,1024,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,8192,0.010663111176755695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,3584,0.087826669216156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,3584,0.0073902226156658595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,10240,0.024157333705160353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,512,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,12288,0.01575644479857551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,1024,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,512,0.009102221992280748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,3072,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,5120,0.00959022260374493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,8192,0.07149600320392184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,16384,0.01919555498494042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,768,0.00544177782204416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,512,0.014496000276671516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,2560,0.014507555299335055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,64,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,64,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,512,0.012757333616415659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,10240,0.01956266661485036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,1024,0.01852799952030182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,8192,0.018363555272420246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,6144,0.015257777439223396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,12288,0.01516177753607432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,1024,0.01276711126168569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,1024,0.02501511077086131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,12288,0.019835554891162448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,64,0.013095999757448832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,3072,0.01295111080010732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,256,0.013065777719020844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,10240,0.016751110553741455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,4096,0.131129781405131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,1024,0.005582222094138463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,1536,0.006251555350091722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,10240,0.015659555792808533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,768,0.016926222377353244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,65536,0.20198222001393637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,2048,0.009699555734793345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,64,0.0018737777653667664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,12288,0.044758223825030856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,1536,0.00666222224632899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,768,0.009346666435400644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,12288,0.5965058008829752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,65536,0.04475111100408766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,2048,0.013560000393125745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,7168,0.014821334017647637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,3072,0.19214399655659994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,512,0.038958221673965454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,3072,0.013760888742076026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,3072,0.09351022375954522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,768,0.017419555120997958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,512,0.0052826665341854095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,64,0.008602666358153025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,1024,0.008999110923873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,6144,0.01108000013563368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,1536,0.012086222569147745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,65536,0.07573333051469591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,5120,0.018292443619834054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,3584,0.0151235560576121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,2560,0.028463111983405218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,3584,0.007272000114123027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,5120,0.00416355538699362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,256,0.013079110946920184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,768,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,256,0.01479111115137736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,1536,0.008741333252853816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,8192,0.011358222199810876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,16384,0.05211288730303446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,3072,0.011366222467687396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,12288,0.03468177715937296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,1024,0.014170666535695394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,32,0.002518222149875429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,7168,0.018571555614471436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,3584,0.06067733632193672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,8192,0.11547466119130452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,768,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,2560,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,128,0.014448000325096978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,16384,0.021015110943052504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,2560,0.007245333658324347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,1024,0.024110222856203716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,3584,0.011400889191362592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,256,0.02637422250376807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,7168,0.056537780496809215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,64,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,3072,0.048885332213507764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,6144,0.02129511038462321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,5120,0.016752888758977253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,2560,0.1227662165959676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,8192,0.01775911119249132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,3072,0.014411555396185981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,7168,0.01737955543729994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,3072,0.01587022178702884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,1536,0.006239111224810283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,512,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,10240,0.01074044406414032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,512,0.018535999788178336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,12288,0.17810489071740043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,10240,0.014767110347747803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,64,0.0018951110541820526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,7168,0.011753777662913004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,3584,0.029440889755884807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,2048,0.017613333132531907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,512,0.0028862222615215513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,2560,0.01147466649611791
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,10240,0.021985777550273474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,768,0.017575111654069688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,65536,0.06200177801979912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,12288,0.11612533198462592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,12288,0.02493422230084737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,4096,0.02312266661061181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,6144,0.005720888988839255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,128,0.004955555415815777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,6144,0.008991999758614434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,1024,0.005917333480384614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,2048,0.0199626667631997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,10240,0.012121777567598553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,3584,0.008996444443861643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,7168,0.01330133361948861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,3072,0.03755377729733785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,512,0.05651022328270806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,256,0.016197333733240765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,128,0.002128888956374592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,512,0.002465777720014254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,65536,0.15841778119405112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,65536,0.2169342173470391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,7168,0.01553244392077128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,768,0.012097777591811286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,2560,0.006618666566080517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,32,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,5120,0.016539555456903245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,16384,0.005956444475385878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,3584,0.004552000098758274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,256,0.0021537778278191886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,1024,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,1024,0.005201777650250329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,256,0.013743110828929476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,2560,0.03254133462905884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,5120,0.015793777174419828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,3584,0.008616888688670265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,4096,0.030300445026821558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,6144,0.13583021693759495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,5120,0.019332443674405415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,256,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,3584,0.02234755622016059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,8192,0.16353244251675075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,2560,0.021038222644064162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,512,0.00254488881263468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,16384,0.027373333772023518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,1536,0.026141333911154006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,10240,0.012159110771285163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,3584,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,7168,0.026029333472251892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,10240,0.0410097771220737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,2048,0.007717333734035492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,2048,0.00902844468752543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,768,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,8192,0.02437244521247016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,5120,0.006960888703664144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,8192,0.031038221385743883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,2048,0.035953776703940496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,1024,0.29195645120408803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,3072,0.042232000165515475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,3072,0.009291555318567488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,3072,0.007289778027269576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,5120,0.008008889026112026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,32,0.0052391112678580815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,2560,0.030128889613681372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,16384,0.005410666680998272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,32,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,16384,0.01687733332316081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,16384,0.025055110454559326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,3584,0.027564444475703772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,2560,0.013841778039932251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,12288,0.014151111245155334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,2560,0.006959999601046245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,12288,0.015490666031837463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,10240,0.026775999201668635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,768,0.006919999917348226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,64,0.0021946666141351065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,10240,0.012667555775907306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,7168,0.11179199483659531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,1536,0.012840000291665396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,256,0.016196444630622864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,65536,1.2546052932739258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,4096,0.006923555499977536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,128,0.016135111451148987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,3072,0.03490133417977227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,3072,0.008983111215962304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,4096,0.04909244510862562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,7168,0.01236088905069563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,4096,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,6144,0.010955555571450127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,512,0.0034471110751231513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,1024,0.008714666797055138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,1024,0.0035182221068276297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,65536,0.14322043789757624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,4096,0.022262222237057153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,256,0.004902222090297275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,256,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,7168,0.009667555491129557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,256,0.0028657778683635923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,1536,0.00777511133088006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,256,0.002199999988079071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,1024,0.005582222094138463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,4096,0.13213600052727595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,7168,0.02256444427702162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,2560,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,16384,0.011752000285519494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,10240,0.009569777382744683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,7168,0.03280266788270738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,5120,0.03983111182848612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,64,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,64,0.003024000053604444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,1536,0.0141004439857271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,2560,0.007518221934636434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,1024,0.0034933334423436057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,2560,0.007393777370452881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,5120,0.013751111096805997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,768,0.013231111069520315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,1024,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,4096,0.012081777883900536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,256,0.03847733471128676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,65536,0.09567111068301731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,10240,0.01682577696111467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,6144,0.5303369098239475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,4096,0.0769875579410129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,7168,0.016657777958446078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,4096,0.023910222782029047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,1536,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,768,0.017342221405771043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,16384,0.026164443956481084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,64,0.002495999965402815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,256,0.021936888496081035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,8192,0.010683555569913654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,5120,0.008652444514963362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,1536,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,12288,0.028138667345046997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,2560,0.07665244738260905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,1024,0.015454222758611044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,5120,0.02470933397610982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,2048,0.036694222026401095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,16384,0.01685066686736213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,6144,0.012132444315486483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,16384,0.3579084343380398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,768,0.005933333188295364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,2560,0.004922666483455234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,6144,0.8543670972188314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,8192,0.01610933409796821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,16384,0.04387022058169047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,7168,0.03807555635770162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,1536,0.005931555396980709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,10240,0.022463111413849726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,65536,0.13448354932996961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,64,0.0021173333128293357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,1536,0.005938666562239329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,64,0.002116444417172008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,2048,0.016833777228991192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,16384,0.02273155583275689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,256,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,768,0.0027031112048361036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,768,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,16384,0.014626666903495789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,2560,0.005196444276306364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,5120,0.023896889554129705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,128,0.002502222235004107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,32,0.00554044461912579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,768,0.0038373333712418876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,5120,0.026135111848513286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,3584,0.07195022371080187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,2048,0.004592000196377437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,10240,0.06328177452087402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,12288,0.03679644399219089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,65536,0.10981688896814983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,128,0.04246755441029867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,65536,6.524265713161892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,768,0.007274666594134436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,768,0.007283555964628856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,1536,0.006706666615274217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,512,0.0028595555987623003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,2560,0.008962666822804345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,6144,0.08736266692479451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,16384,0.31041155921088326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,5120,0.1470257838567098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,7168,0.009306666751702627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,12288,0.03512800070974562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,8192,0.010045333041085137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,12288,0.37096712324354386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,8192,0.010377777947319878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,1024,0.01629955569903056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,1024,0.020579554968410067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,768,0.028781334559122723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,512,0.0030053332448005676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,512,0.011363555159833698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,2560,0.06300799714194404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,8192,0.15044711695777044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,7168,0.027582221561008032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,16384,0.04868533213933309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,256,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,8192,0.20244267251756456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,16384,0.03526844580968221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,5120,0.056636446052127414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,12288,0.05932088692982992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,6144,0.010672888822025724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,8192,0.10156266556845771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,128,0.002507555608948072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,2560,0.007284444239404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,2048,0.004587555511130227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,10240,0.09928800000084771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,8192,0.013367999758985309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,10240,0.011711111499203576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,128,0.009755555954244401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,6144,0.03327911098798116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,2048,0.00796088907453749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,2560,0.004852444347408083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,10240,0.026327111654811438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,256,0.002206222257680363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,128,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,16384,0.011708444191349877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,1536,0.039487110243903265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,4096,0.009833777944246927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,8192,0.041647109720442034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,1536,0.005514666852023866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,10240,0.017172444197866652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,65536,0.2155848873986138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,65536,0.008071111308203803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,1536,0.12027466297149658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,3584,0.009732444253232744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,256,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,768,0.0027822222974565295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,64,0.0041848888827694785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,2560,0.006587555425034628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,7168,0.009648888475365108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,512,0.027106665902667578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,8192,0.017128888103697035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,12288,0.014107555150985718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,10240,0.01998133295112186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,128,0.004944889081848992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,768,0.005943111247486538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,4096,0.008182221816645728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,64,0.005698666804366642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,10240,0.0939626693725586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,2048,0.016922665966881644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,256,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,64,0.008654221892356873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,7168,0.012072888513406118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,5120,0.009321778184837764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,512,0.006897777732875612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,64,0.0060159998635451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,2048,0.006696888970004187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,128,0.007348444726732042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,3584,0.00794400026400884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,32,0.0069884442620807225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,16384,0.36609776814778644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,256,0.006695111178689533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,768,0.003481777591837777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,8192,0.014517333772447375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,512,0.0025333333760499954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,128,0.004856889032655292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,1536,0.006250666661394968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,32,0.0018968888454967076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,10240,0.006603555546866522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,2048,0.019163555569118924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,4096,0.014740443891949125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,16384,0.039284444517559476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,2560,0.01384800010257297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,2048,0.056582219070858426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,3584,0.007650666766696506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,16384,0.025153777665562097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,4096,0.007679111427730984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,1024,0.006263999889294307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,2048,0.008776889079146916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,6144,0.009305777649084726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,16384,0.033125334315829806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,2048,0.02300888962215847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,5120,0.029398222764333088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,1536,0.056904892126719155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,256,0.014097778333557976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,10240,0.011353777514563667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,768,0.00573511090543535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,2560,0.02132622235351139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,768,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,768,0.012401777837011548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,10240,0.008641777767075432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,4096,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,3584,0.08726311392254299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,5120,0.009701333112186855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,512,0.003048000029391713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,64,0.004864888886610667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,3584,0.009659555223253038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,12288,0.009124444590674506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,65536,0.3222160074445936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,2048,0.011000889043013254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,12288,0.04091466797722711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,2048,0.0041893331540955436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,5120,0.01681866745154063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,256,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,3584,0.04038666685422262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,5120,0.00943377779589759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,256,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,64,0.005201777650250329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,1536,0.015818667080667283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,12288,0.023318222827381555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,1536,0.010536889235178629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,768,0.09653066926532322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,1024,0.0031822222388452957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,4096,0.01128355579243766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,12288,0.035242666800816856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,4096,0.014103111293580798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,65536,0.09601600302590264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,3072,0.007994666695594788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,3072,0.05056266652213203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,65536,0.08566666973961724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,768,0.00554577757914861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,16384,0.022631999519136217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,1024,0.01314133322901196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,12288,0.0840275552537706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,512,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,6144,0.025469332933425903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,64,0.002206222257680363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,256,0.00591822216908137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,1024,0.019626667102177937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,256,0.014839111102951897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,3584,0.008683555656009251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,2048,0.026485333840052288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,16384,0.1694115532769097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,256,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,65536,0.32570311758253306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,7168,0.05097066693835788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,12288,0.030516445636749268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,512,0.004371555729044808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,768,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,3584,0.007963555554548899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,256,0.010394666757848527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,2560,0.007152000235186682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,1536,0.016531555189026725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,65536,0.11478843953874375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,10240,0.01625155574745602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,512,0.00941955546538035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,6144,0.2678533395131429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,768,0.006701333241330252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,512,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,2048,0.006543111056089401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,6144,0.018406222263971966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,512,0.008482666479216682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,7168,0.009930666950013902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,6144,0.0110880004035102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,16384,0.040008889304267034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,3072,0.02219466699494256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,1536,0.0232924438185162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,2560,0.006960888703664144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,512,0.01311822235584259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,65536,0.009346666435400644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,65536,0.3092613220214844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,8192,0.025423111187087163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,6144,0.018226666582955253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,5120,0.027219555444187585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,5120,0.21577066845364043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,64,0.018575999471876357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,768,0.005535111245181825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,2048,0.012478222449620565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,5120,0.010513777534166971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,2048,0.006615110983451207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,4096,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,256,0.005053333524200651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,16384,0.02555822167131636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,768,0.02369244396686554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,16384,0.07165600193871392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,768,0.005576000031497743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,10240,0.024382222029897902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,5120,0.03858400053448147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,768,0.017082666357358296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,1024,0.005964444329341252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,65536,0.051741331815719604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,1024,0.002903999967707528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,4096,0.017687999539905123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,256,0.006629333313968446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,10240,0.02646133303642273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,10240,0.021243555678261652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,3072,0.09160088830524021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,3072,0.21494844224717882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,65536,0.07782666550742255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,256,0.014097778333557976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,8192,0.02096533278624217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,2560,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,2048,0.006237333433495627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,6144,0.018550222118695576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,12288,0.019243554936514962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,4096,0.012413333687517377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,10240,0.02725511127048069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,768,0.010866666833559671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,64,0.004972444640265571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,65536,0.05442844496832954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,2560,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,10240,0.017095110482639737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,1536,0.003547555456558863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,12288,0.018901333212852478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,7168,0.013346666263209449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,5120,0.008256889051861232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,3584,0.008997333546479544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,65536,0.07133866680992974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,5120,0.008618666893906063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,7168,0.03850311040878296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,1536,0.007322666545708974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,12288,0.02665422194533878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,768,0.0028311111446883944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,65536,0.1910026603274875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,1536,0.007605333295133378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,512,0.023207111491097346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,65536,0.15790666474236384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,2560,0.015971556305885315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,2560,0.02947910957866245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,65536,0.15044177903069392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,768,0.010284444524182213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,64,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,10240,0.15139910909864637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,512,0.004868444469239977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,8192,0.04725689027044508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,6144,0.015655999382336933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,3584,0.01683644453684489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,4096,0.007684444387753804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,65536,0.08283199866612752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,512,0.004993777722120285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,2560,0.006441777778996362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,2560,0.004936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,3072,0.01791555517249637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,16384,0.025742222865422566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,256,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,256,0.006042666733264923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,1536,0.012121777567598553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,1024,0.02743911080890232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,256,0.015184889237085978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,4096,0.009008888569143083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,5120,0.04474755459361606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,65536,0.09015466769536336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,7168,0.029417776399188574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,4096,0.023401778605249193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,5120,0.010075555907355415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,128,0.004929777648713854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,64,0.0021377777059872947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,65536,0.044136001004113086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,6144,0.004874666531880696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,6144,0.022154667311244543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,7168,0.2243839899698893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,3584,0.007262222468852997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,1536,0.015970667203267414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,3584,0.007285333342022366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,7168,0.014766222900814481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,6144,0.0425217780801985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,12288,0.004917333523432414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,16384,0.037471999724706016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,512,0.014149333039919535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,4096,0.0052044445441828836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,6144,0.009002666506502364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,3584,0.013312000367376538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,512,0.002577777744995223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,768,0.01742044422361586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,7168,0.030055999755859375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,10240,0.013516444298956128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,7168,0.010071111222108206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,7168,0.010267555713653564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,8192,0.012154666913880242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,8192,0.032765332195493914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,3584,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,4096,0.010911111202504901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,512,0.013594666288958656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,2048,0.039352890517976545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,65536,0.39603643947177464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,10240,0.016554666890038382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,256,0.004889777965015835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,4096,0.1090666651725769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,6144,0.013276444541083442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,16384,0.15964355733659533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,7168,0.05184266633457608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,64,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,256,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,64,0.02404088940885332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,65536,0.07508355379104614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,768,0.004634666774008009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,768,0.002485333424475458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,10240,0.03390489021937052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,128,0.013786666923099093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,12288,0.04059377643797133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,6144,0.015042665931913586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,2048,0.01347644461525811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,8192,0.016898666818936665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,12288,0.03523733218510946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,6144,0.018945778409639995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,256,0.004253333227501975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,128,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,256,0.009024889104896123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,128,0.0021662221600612006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,1536,0.006311999840868845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,64,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,7168,0.011920000116030375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,8192,0.013556444810496436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,3072,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,10240,0.12984355290730795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,7168,0.009672889278994666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,65536,0.07773422532611422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,12288,0.017441777719391715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,768,0.015531554818153381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,6144,0.004581333448489507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,10240,0.04899911085764567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,8192,0.038182222180896334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,128,0.02125955621401469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,768,0.014104889498816596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,3584,0.007956444389290279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,4096,0.009999111294746399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,3584,0.04723466767205132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,768,0.0032097777972618737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,2048,0.1144604418012831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,3584,0.020441777176327176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,8192,0.01814133259985182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,64,0.011754666765530905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,768,0.014456889695591398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,3072,0.016545777519543965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,1536,0.01806400054030948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,4096,0.007643555601437886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,2048,0.004578666554556953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,1536,0.01477866702609592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,16384,0.045957333511776395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,2560,0.006491555521885554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,8192,0.018927110566033255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,128,0.015473778049151102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,32,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,2560,0.004208000169859992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,2048,0.024127110838890076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,8192,0.00942755573325687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,16384,0.23119023111131456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,65536,4.1828969319661455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,6144,0.01460533340771993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,12288,0.07858755853441027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,16384,0.02201333310869005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,1024,0.0199635558658176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,768,0.007265778051482306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,1024,0.016466667254765827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,768,0.018523555662896898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,3072,0.01756266587310367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,10240,0.016576889488432143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,64,0.006375999914275275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,2048,0.07262577613194783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,512,0.0135057775510682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,65536,0.1564320060941908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,2560,0.021754667162895203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,16384,0.015443555182880826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,65536,0.05992977486716377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,3072,0.07564266522725423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,2048,0.008640000389681922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,2048,0.006285333385070165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,1536,0.004259555704063839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,4096,0.085189335876041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,512,0.002522666628162066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,12288,0.04269955555597941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,3072,0.03287200133005778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,768,0.0035182221068276297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,3584,0.012028444144460889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,128,0.014470222923490735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,2560,0.009357333183288574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,1536,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,32,0.0018728888697094387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,8192,0.011507555842399597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,7168,0.009648000200589498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,128,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,5120,0.013777777552604675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,65536,0.09746844238705105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,10240,0.1841733323203193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,65536,0.8759831322564019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,4096,0.011157333023018308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,2048,0.012404444317022959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,3072,0.008630221916569604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,3072,0.018616000811258953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,4096,0.04780888888570997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,2048,0.01255555533700519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,512,0.05967289209365845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,128,0.004929777648713854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,8192,0.015444444285498725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,1024,0.009100444614887238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,4096,0.1231715546713935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,32,0.019557333654827543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,5120,0.02313422163327535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,4096,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,6144,0.013780444032616086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,32,0.008647999829716152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,2560,0.019262222780121695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,2560,0.01794933279355367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,768,0.017521777086787753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,65536,0.07644355297088623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,2560,0.008984000318580205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,2048,0.01238933371173011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,12288,0.10948266585667928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,8192,0.016913778252071805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,7168,0.009986666341622671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,768,0.0028133332315418455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,3584,0.011062222222487131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,2560,0.00833511104186376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,65536,0.7596035533481174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,10240,0.03158933255407546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,5120,0.05423911081420051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,10240,0.006262222097979651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,12288,0.10950844155417548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,6144,0.00592533333433999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,6144,0.11769778198666042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,1536,0.005435555759403441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,3584,0.09837955236434937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,8192,0.01990133358372582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,128,0.007706666986147563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,5120,0.010392000277837118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,64,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,10240,0.016551999582184684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,16384,0.040844443771574236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,256,0.01443377799457974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,65536,0.0604355533917745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,12288,0.017839999662505258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,128,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,3584,0.011738667057620155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,16384,0.04179822074042427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,16384,0.059024002816942006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,8192,0.026744888888465032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,10240,0.01627111103799608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,64,0.005321777943107817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,3584,0.020055999358495075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,64,0.002536000063021978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,768,0.0041991112132867174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,32,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,5120,0.012895111408498553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,1536,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,256,0.0024702221983008915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,12288,0.018814222680197824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,3072,0.00759733302725686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,128,0.005167110926575131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,8192,0.02091022166940901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,768,0.003148444410827425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,4096,0.007941332956155142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,65536,0.1194346613354153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,10240,0.07723466555277507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,2560,0.013693333499961428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,7168,0.14601155122121176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,512,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,2048,0.012081777883900536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,3584,0.011959999799728394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,5120,0.004583111239804162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,6144,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,256,0.00286666676402092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,768,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,16384,0.020982222424613103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,2560,0.01854933301607768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,65536,0.08202399810155232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,1536,0.024804444776640996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,2048,0.0138568886452251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,6144,0.013722666435771518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,16384,0.10081955459382798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,8192,0.01071999967098236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,5120,0.008678221868144141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,128,0.007381333245171442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,7168,0.029713776376512315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,2048,0.017849778135617573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,768,0.0028613333900769553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,8192,0.16243910789489746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,65536,0.08468266990449692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,4096,0.00906755526860555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,3072,0.004646222210592694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,128,0.008052444292439355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,16384,0.02281155520015293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,4096,0.06165422333611382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,10240,0.030480888154771592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,32,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,1536,0.017567111386193168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,4096,0.020762667059898376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,2560,0.19408622052934435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,8192,0.02792888879776001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,4096,0.007956444389290279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,4096,0.016873778568373788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,7168,0.020350222786267597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,1024,0.022713777091768052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,12288,0.012812444733248817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,2560,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,3072,0.01887022289964888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,5120,0.030722667773564655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,65536,0.08112711376614042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,12288,0.04208888941340976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,3072,0.0605600012673272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,2560,0.044044444958368935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,3072,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,12288,0.01854044364558326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,10240,0.1115857760111491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,6144,0.03849155704180399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,10240,0.1464328898323907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,256,0.004867555366622077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,3072,0.010204444328943888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,768,0.02847644355561998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,7168,0.028612444798151653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,8192,0.03172355559137132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,512,0.007937777373525832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,8192,0.01312711089849472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,32,0.006288888967699475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,6144,0.008999110923873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,1024,0.006061333335108227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,3072,0.030861334668265447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,3072,0.009076444639099969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,6144,0.03942844602796767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,6144,0.009633777870072259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,32,0.001810666587617662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,12288,0.13150755564371744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,16384,0.2799866729312473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,1536,0.014531556102964612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,8192,0.01164177805185318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,10240,0.1922382248772515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,12288,0.024430221981472437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,65536,0.07652799950705634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,65536,0.08226222462124295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,65536,0.05547822184032864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,768,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,8192,0.07598400115966797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,128,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,65536,0.13916444778442383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,2048,0.006228444476922353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,2560,0.03392533461252848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,2560,0.0837528904279073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,3584,0.07644710938135783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,2048,0.015127999915017022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,16384,0.014045332868893942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,16384,0.018214222457673814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,7168,0.022896000080638464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,10240,0.04172977805137634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,1024,0.006279110908508301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,128,0.0020986667109860317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,16384,0.08353510830137466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,6144,0.019967110620604623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,128,0.0022444443570242985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,768,0.003938666648334927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,512,0.005577777822812398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,3584,0.01684799955950843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,8192,0.012999999854299756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,2560,0.006621333460013072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,1536,0.011692444483439127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,1536,0.0035217776894569397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,5120,0.08715911044014825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,6144,0.021617778473430212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,65536,5.18560791015625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,768,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,7168,0.010374222364690568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,7168,0.016751999656359356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,7168,0.017808889349301655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,1536,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,32,0.008546666966544257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,6144,0.008349333372380998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,12288,0.007401777638329401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,2560,0.08350133233600193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,65536,0.07853511306974623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,4096,0.02332177758216858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,12288,0.011365333365069495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,65536,0.08152800136142306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,768,0.007296000089910295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,5120,0.009359111388524374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,4096,0.008344888687133789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,3584,0.07126222054163615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,10240,0.045040001471837364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,7168,0.015572445260153877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,256,0.003863111138343811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,256,0.0025191110455327565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,64,0.005486222190989389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,5120,0.009663110805882348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,128,0.02829244401719835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,3584,0.007975111405054728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,2560,0.006940444310506185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,10240,0.01204355557759603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,3584,0.02408711115519206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,8192,0.01629688839117686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,1024,0.009679999616410997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,65536,0.19134844674004448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,3584,0.007323555648326874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,3072,0.015771556231710646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,1024,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,3584,0.020227554771635268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,12288,0.028192000256644353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,768,0.017481777403089736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,10240,0.07973066965738933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,4096,0.007842666572994655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,256,0.002177777803606457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,3072,0.11887644396887885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,768,0.014767110347747803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,128,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,768,0.01240533341964086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,1024,0.016183111402723525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,65536,0.15080800321367052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,4096,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,16384,0.061476442548963756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,2048,0.022758222288555566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,6144,0.01719999975628323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,3584,0.011861333416567909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,16384,0.14947111076778835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,5120,0.3996275530921088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,64,0.0025155554629034465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,7168,0.009648000200589498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,2560,0.008403555386596257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,12288,0.08366933133867051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,32,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,6144,0.013304000099500021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,12288,0.028359999259312946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,5120,0.01418133411142561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,12288,0.1206702258851793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,512,0.013362666798962487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,64,0.015799999237060547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,512,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,16384,0.05173244410090976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,1536,0.003896000070704354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,64,0.0018639999131361644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,8192,0.008635555704434713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,3072,0.02605688903066847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,2560,0.007608888877762689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,2560,0.01959822244114346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,1536,0.005912000106440649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,8192,0.02010222276051839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,2560,0.016155555844306946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,7168,0.009588444398509132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,65536,0.0787680016623603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,8192,0.029513776302337646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,256,0.002529777793420686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,16384,0.025423111187087163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,512,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,3584,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,6144,0.017400888933075797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,2560,0.006329777754015393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,16384,0.05971022446950277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,2560,0.014819555812411837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,12288,0.055086221959855824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,128,0.026022222306993272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,16384,0.03960444529851278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,128,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,2560,0.021433777279324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,1024,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,7168,0.012620444099108378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,4096,0.022858666049109563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,128,0.005107555538415909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,5120,0.008348444269763099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,12288,0.016738666428460013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,10240,0.020241777102152508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,128,0.014118222726715935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,2560,0.015201777219772339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,8192,0.048376888036727905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,16384,0.04723822077115377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,8192,0.025195555554495916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,64,0.007496000164084964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,65536,0.21082577440473768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,32,0.0019742221468024785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,10240,0.047804445028305054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,5120,0.008233778178691864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,4096,0.023057777020666335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,7168,0.020853334003024634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,12288,0.022618666291236877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,64,0.0028035555862718155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,12288,0.02774488925933838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,3072,0.008400888906584846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,65536,0.08519022332297431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,10240,0.018637332651350234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,65536,0.10992711120181614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,128,0.013776889277829064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,256,0.0029680000411139596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,5120,0.013034666577974955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,2560,0.00835111074977451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,32,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,7168,0.022991999983787537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,2560,0.06811466481950548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,16384,0.030582222673628066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,2048,0.01556711064444648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,6144,0.0047644442982143825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,2048,0.004874666531880696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,2048,0.014128888646761576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,2048,0.060080885887145996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,128,0.013787555197874704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,128,0.01477599971824222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,5120,0.027126222848892212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,2560,0.007568000091446771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,2560,0.006890666567616993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,8192,0.01498222185505761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,7168,0.02029244436158074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,768,0.003133333391613431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,512,0.008279110822412703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,32,0.005055111315515307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,5120,0.030524445904625788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,7168,0.011225777367750803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,128,0.03909777932696872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,512,0.005173333403136996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,256,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,8192,0.010034666293197209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,8192,0.016385777129067313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,7168,0.0148053334818946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,64,0.007954667011896769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,64,0.0021608887861172357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,2048,0.04967022273275587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,2048,0.01921422282854716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,12288,0.014112888111008538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,7168,0.017858665850427415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,6144,0.010363555616802638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,64,0.002102222293615341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,10240,0.012044444680213928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,2560,0.004943999979231092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,1024,0.0075937774446275495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,2048,0.010796444283591377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,4096,0.007630222373538547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,2560,0.025424000289705064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,512,0.023987554841571387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,6144,0.014592889282438489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,768,0.0053119998839166425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,3584,0.008455110920800103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,2048,0.006531555619504716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,768,0.01683644453684489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,2560,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,6144,0.012549333274364471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,4096,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,768,0.006285333385070165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,12288,0.026426666312747534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,10240,0.057144886917538114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,1536,0.012582222620646158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,32,0.0069475554757648045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,7168,0.013912000589900546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,4096,0.015885333220163982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,8192,0.04612444506751167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,2048,0.01853511068556044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,3584,0.007251555720965068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,2048,0.042839109897613525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,10240,0.027512888113657635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,2048,0.0311173333062066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,6144,0.20343644089168975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,10240,0.02539199921819899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,5120,0.009388444324334463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,128,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,128,0.005984000033802456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,4096,0.020230222079488967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,5120,0.018992000155978732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,1024,0.009322666459613377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,5120,0.02365600069363912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,5120,0.011843555503421359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,16384,0.14731378025478786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,512,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,3072,0.018236445056067575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,768,0.0240000006225374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,4096,0.014119999276267158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,64,0.005924444645643234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,256,0.004927111168702443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,2048,0.006573333508438534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,6144,0.0064080001579390625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,65536,0.1933493349287245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,5120,0.020015999674797058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,2048,0.006239111224810283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,6144,0.008968888885445064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,6144,0.06442577971352471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,2048,0.007278222176763747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,5120,0.029577778445349798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,12288,0.5686773194207085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,1536,0.015848888291252982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,512,0.07357688744862874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,65536,0.09328089157740276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,65536,0.18866933716668022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,8192,0.18598311477237275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,65536,0.03573777940538194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,7168,0.011074666347768573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,64,0.005569777968857024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,64,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,5120,0.034655110703574285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,10240,0.013616888887352414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,12288,0.045913779073291354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,5120,0.010279110736317104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,12288,0.01941244469748603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,128,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,3072,0.010038221875826517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,5120,0.02076088885466258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,10240,0.03052355514632331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,12288,0.01921777758333418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,10240,0.10770489109887017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,32,0.01927199959754944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,3072,0.02495199938615163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,4096,0.04105866617626614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,3072,0.007623111208279927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,768,0.011109333071443768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,256,0.002469333302643564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,1024,0.00592533333433999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,128,0.0021359999146726397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,64,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,1536,0.007755555212497711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,1024,0.004913777940803104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,4096,0.030173334810468886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,2560,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,256,0.004949333353175057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,512,0.0063493334584765965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,2048,0.016223111086421542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,3584,0.02889066603448656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,7168,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,12288,0.12720978260040283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,10240,0.05357955561743843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,256,0.020038222273190815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,128,0.005695999910434087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,4096,0.22159022755093047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,8192,0.02460622290770213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,768,0.016148444679048326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,7168,0.01829866733815935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,1024,0.005622222191757626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,3072,0.09231911102930705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,6144,0.015834665960735746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,12288,0.028303109937243994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,16384,0.023412444525294836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,12288,0.052700443400277026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,12288,0.021595555875036452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,3072,0.009637333452701569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,3072,0.09537155760659112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,768,0.006657777975002925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,3584,0.018232888645595975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,4096,0.008967999782827165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,5120,0.009577777650621202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,6144,0.021726222501860723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,1536,0.009531555904282464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,3584,0.016161777906947665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,3072,0.012058667010731168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,2560,0.011797332929240333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,32,0.005884444548024072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,8192,0.014204444156752693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,2560,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,128,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,4096,0.18316711319817436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,6144,0.023377777801619634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,512,0.030508445368872747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,4096,0.026172444224357605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,2560,0.020920889245139226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,8192,0.019309333629078336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,64,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,10240,0.04683199856016371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,256,0.0021884443445338146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,8192,0.010778666370444827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,2048,0.010220444036854638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,4096,0.01744622157679664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,65536,0.29841865433586967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,1536,0.014117333624098035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,12288,0.06316088967853122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,12288,0.022621333599090576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,512,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,1024,0.003160888950030009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,3584,0.0076044441925154785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,5120,0.004967111266321606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,512,0.006642666541867786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,64,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,512,0.0031795555518733132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,256,0.005251555393139522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,256,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,512,0.01353422221210268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,3584,0.010315555665228102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,2048,0.006952889263629913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,7168,0.057526224189334445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,768,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,6144,0.01927199959754944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,6144,0.005882666756709416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,10240,0.012061333490742577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,1536,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,16384,0.019706666469573975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,1536,0.029314666986465454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,3072,0.011368888947698804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,6144,0.018183110488785636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,768,0.004953777624501122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,10240,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,2048,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,7168,0.012454222473833295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,2048,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,3072,0.010653333531485664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,4096,0.010033778018421596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,256,0.004936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,16384,0.027813331948386297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,6144,0.00903733323017756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,3584,0.004566222015354368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,2560,0.0122079998254776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,4096,0.18091111712985566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,12288,0.020629333125220407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,128,0.032368001010682844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,1024,0.016193777322769165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,768,0.03859466645452712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,32,0.018343110879262287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,512,0.005375111268626318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,8192,0.02438933319515652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,3072,0.02141155633661482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,16384,0.09956444634331597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,3584,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,1024,0.003900444342030419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,16384,0.008029333419269985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,1536,0.037696000602510236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,128,0.001810666587617662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,2048,0.006984000404675801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,12288,0.17866667111714682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,256,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,128,0.02755733331044515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,10240,0.05742488967047798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,16384,0.09990133179558648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,10240,0.012063110868136087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,128,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,3584,0.007998222278224098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,7168,0.01033333357837465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,10240,0.012077333198653327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,10240,0.02035555574629042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,4096,0.009490666290124258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,65536,0.26201777988009983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,65536,0.04445688923199972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,1536,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,128,0.00498488876554701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,8192,0.015849777393870883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,1536,0.008085333638721043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,65536,0.06243199772304959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,8192,0.16338311301337347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,7168,0.04790488878885905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,2048,0.018060444129837882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,8192,0.016173332929611206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,2560,0.006956444846259222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,6144,0.12625332673390707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,12288,0.0148053334818946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,1536,0.006263111200597551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,256,0.023638221952650283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,1536,0.009392889009581672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,7168,0.0336497790283627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,10240,0.008575999902354347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,65536,0.18034666114383271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,2560,0.004228444563017951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,6144,0.04288799895180596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,16384,0.08515199687745835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,4096,0.028573334217071533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,1024,0.03509511219130622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,4096,0.021020443903075323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,3072,0.027561777167850073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,1536,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,3584,0.010300444232092964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,5120,0.02905333373281691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,7168,0.010220444036854638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,2560,0.040842665566338435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,4096,0.009191999832789103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,3584,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,12288,0.01721155477894677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,1024,0.003908444609906939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,2560,0.010284444524182213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,2560,0.08729422092437744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,12288,0.01310222182008955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,128,0.01512977812025282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,12288,0.03040177623430888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,2560,0.009647111097971598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,6144,0.008965333302815756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,12288,0.049581332339180835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,128,0.002209777840309673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,12288,0.03700977894994948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,65536,0.15464088651869032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,10240,0.016146666473812528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,6144,0.009503111243247986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,128,0.01812799937195248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,3072,0.005210666606823604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,1024,0.01586577792962392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,8192,0.00591822216908137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,1024,0.011518222590287527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,1536,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,1536,0.018184888694021437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,7168,0.009995555712117089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,256,0.20257511403825548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,32,0.0028151110228565005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,3584,0.007782222496138678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,2048,0.014012444350454541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,6144,0.009325332939624786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,3072,0.09986844327714707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,2048,0.009009777671760982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,4096,0.008789333204428354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,6144,0.026888888743188646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,3584,0.007248000138335758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,3584,0.007668444679843054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,1536,0.04941244588957893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,5120,0.018301332990328472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,8192,0.028071112102932397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,10240,0.013281777501106262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,64,0.004576000074545543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,7168,0.010299555957317352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,1536,0.026379555463790894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,3072,0.007296889192528195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,128,0.00692622239391009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,5120,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,768,0.0028968888024489083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,6144,0.011477332976129321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,65536,0.07722044653362699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,512,0.00656088896923595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,256,0.013061333033773633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,256,0.007956444389290279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,32,0.0029102222373088202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,10240,0.024658666716681585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,1536,0.015498666299713982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,128,0.014137778017255994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,7168,0.02698844505680932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,2560,0.029489778810077246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,10240,0.03832799858517117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,2048,0.006622222148709827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,5120,0.013807999591032663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,7168,0.013979555832015144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,1024,0.00589155571328269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,65536,0.0806844433148702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,3584,0.031349334451887347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,5120,0.010375111467308467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,12288,0.09191377957661946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,6144,0.02652888827853733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,2048,0.010026666853162978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,65536,0.1466017829047309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,1024,0.005945777727497949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,6144,0.010223111344708337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,16384,0.2844853401184082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,1536,0.012501333322789935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,2048,0.008229333493444655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,3584,0.007998222278224098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,10240,0.3135502338409424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,3584,0.041971554358800255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,256,0.056905779573652476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,12288,0.03480444351832072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,7168,0.03735466798146566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,768,0.005601777798599667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,2560,0.006566222343179915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,10240,0.07659466399086846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,3072,0.1936213307910495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,8192,0.02443733314673106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,3584,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,5120,0.01350933313369751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,6144,0.012875555290116204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,8192,0.011924444801277585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,2048,0.010833777487277985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,5120,0.02270577847957611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,512,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,5120,0.01779733267095354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,12288,0.013775111072593264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,1024,0.025084444218211707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,1536,0.022962666220135156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,512,0.022641777992248535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,512,0.015458666616015963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,128,0.0025066667132907445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,7168,0.03404622276624044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,3584,0.077729774845971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,128,0.016439111696349252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,6144,0.06896622313393487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,5120,0.008655110994974772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,7168,0.01442488862408532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,12288,0.044024000565210976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,2560,0.006727111008432176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,64,0.0516480008761088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,1024,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,8192,0.011015111373530494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,4096,0.019189332922299702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,128,0.014344000154071383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,512,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,512,0.0038533334930737815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,8192,0.027580445011456806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,2560,0.008974221845467886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,8192,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,5120,0.012741333080662621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,6144,0.03536355495452881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,128,0.0021253333737452826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,2048,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,6144,0.011104889214038849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,6144,0.009391110804345872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,768,0.035191112094455294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,64,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,3584,0.013204444613721637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,12288,0.04738755689726936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,8192,0.18110044797261557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,4096,0.00792266676823298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,10240,0.020970667401949566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,16384,0.023588443795839947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,5120,0.019938665959570143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,32,0.0019084444890419643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,3072,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,32,0.00216977774269051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,2048,0.031242665317323472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,256,0.19019467300838896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,65536,0.10250400172339545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,768,0.009576000273227692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,10240,0.019148444135983784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,10240,0.03010400136311849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,1536,0.020833777056799993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,6144,0.011149333582984077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,10240,0.07990044355392456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,5120,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,7168,0.017305778132544625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,16384,0.0225262227985594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,4096,0.00425511101881663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,1024,0.013460444079505073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,65536,0.008610666626029545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,32,0.007284444239404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,768,0.014783999986118741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,512,0.007442666424645319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,128,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,32,0.006216889040337668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,4096,0.02034222251839108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,1536,0.03184444374508328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,5120,0.05137866735458374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,256,0.0022728888110982045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,6144,0.010837333069907295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,4096,0.0406622224383884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,6144,0.017557332913080852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,12288,0.022635555929607813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,3584,1.1062471601698134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,10240,0.03863199883037143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,16384,0.040491554472181536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,4096,0.009336888790130615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,2048,0.004507555729813046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,2048,0.009896888501114314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,3584,0.1497973336113824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,7168,4.693943023681641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,4096,0.025251555773946974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,512,0.0041626666982968645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,1024,0.008336000144481659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,5120,0.04931199881765577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,65536,0.07658311393525866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,3584,0.08037600252363417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,4096,0.022802667485343084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,65536,0.00795999997191959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,65536,0.09394311242633396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,512,0.013410666750537025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,1536,0.006582222051090664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,3584,0.028047998746236164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,4096,0.0057964444988303715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,7168,0.027810666296217177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,7168,0.04074222180578444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,2048,0.05700800153944227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,256,0.007714666426181793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,3584,0.02719377809100681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,32,0.0058782220714622075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,7168,0.01549600064754486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,64,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,768,0.00629333323902554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,65536,0.05366222063700358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,64,0.0024764444679021835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,256,0.020999110407299463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,768,0.018156444032986958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,256,0.028419554233551025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,768,0.005607999861240387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,2048,0.018228444788191054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,4096,0.00867022242810991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,4096,0.010401777923107147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,256,0.027296889159414504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,128,0.018805333309703402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,10240,0.0830942193667094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,768,0.0028906667398081887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,12288,0.08437155352698432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,1536,0.017400888933075797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,5120,0.012088889049159156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,64,0.006612444503439798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,128,0.002455111179086897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,10240,0.0185208883550432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,65536,0.5773733456929525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,1536,0.01739288866519928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,2560,0.017103999853134155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,256,0.014094221923086377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,64,0.00519466648499171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,512,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,12288,0.01994044416480594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,7168,0.19775466124216715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,4096,0.028079999817742243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,8192,0.019496889577971566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,2048,0.007163555257850223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,12288,0.07572888665729098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,12288,0.020922667450375024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,10240,0.18883821699354383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,5120,0.016241777274343703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,1024,0.0029075555503368378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,1024,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,10240,0.23687020937601724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,5120,0.03433777888615926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,7168,0.009702222214804756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,65536,1.5734986199273004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,6144,0.0240657776594162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,7168,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,1024,0.01033155537313885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,256,0.0024871110088295406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,16384,0.04545955525504219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,256,0.004982222285535601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,128,0.006598222172922558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,128,0.007985778152942657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,64,0.006609777609507243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,1024,0.12064711252848308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,64,0.0019217777169413036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,5120,0.018579555882347953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,64,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,65536,0.08505244387520684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,7168,0.010325333310498131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,5120,0.02604355580276913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,65536,0.12451644738515218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,32,0.0028684443483750024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,12288,0.005383111122581694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,2560,0.009008888569143083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,2560,0.0677679975827535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,3584,0.007278222176763747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,3584,0.028449777099821303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,16384,0.11034666829639012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,2048,0.010377777947319878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,7168,0.014567111929257711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,2048,0.005626666463083691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,2560,0.006643555644485686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,1024,0.008440888590282863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,128,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,4096,0.007660444411966536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,3072,0.12444710731506348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,3584,0.007306666837798224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,1536,0.004930666751331753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,512,0.02888800038231744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,1536,0.005646222167544895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,256,0.004905777672926585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,1536,0.0038933331767717996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,768,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,1536,0.040644443697399564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,128,0.00565244464410676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,4096,0.011054221954610614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,2048,0.00904977818330129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,768,0.01445777714252472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,65536,0.05414399835798475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,1024,0.07234755489561293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,128,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,7168,0.03730666637420654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,3584,0.006603555546866522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,256,0.0022275555464956495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,128,0.00796888851457172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,10240,0.027134221461084154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,10240,0.06781511174307929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,7168,0.010036444498433007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,768,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,256,0.01481599940194024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,2560,0.004960888789759742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,12288,0.014087110757827759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,6144,0.009024000002278222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,12288,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,2048,0.006577777779764599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,6144,0.05936355723275078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,768,0.005530666559934616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,10240,0.021606221795082092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,7168,0.0528053343296051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,256,0.01444355481200748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,512,0.0031866667171319327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,2048,0.006574222197135289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,128,0.011431111229790581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,65536,0.5774319966634115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,256,0.01613333324591319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,65536,0.23769156138102213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,1536,0.012489777472284106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,512,0.006638222270541721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,3584,0.005931555396980709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,1536,0.02199555602338579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,64,0.002144888871245914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,512,0.004569777597983678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,16384,0.015453333655993143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,2560,0.012092444631788464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,10240,0.01183199965291553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,6144,0.009319999979601966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,512,0.005938666562239329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,128,0.014504000544548035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,1536,0.006331555545330048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,512,0.015173332558737861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,768,0.007631111476156447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,128,0.027052443888452318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,768,0.01754577789041731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,5120,0.009240888886981541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,4096,0.005712000032265981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,4096,0.014352889524565803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,768,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,1024,0.00562755556570159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,768,0.005953777581453323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,2048,0.00977955593003167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,2048,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,12288,0.01479466590616438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,128,0.0021644443687465456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,65536,0.056582219070858426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,256,0.01715377800994449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,5120,0.021056888831986323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,10240,0.08220889170964558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,256,0.002510222295920054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,128,0.18539022074805367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,1536,0.005432888865470886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,10240,0.01039644413524204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,768,65536,0.07946755488713582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,12288,0.017588444881969027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,512,0.013096888860066732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,4096,0.007677333222495184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,7168,0.010983111129866706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,768,0.017557332913080852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,3072,0.08605600065655178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,10240,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,3584,0.1196240054236518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,7168,0.035206221871905856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,3584,0.017948443690935772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,6144,0.013878222141000958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,16384,0.025217778152889673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,16384,0.02149155570401086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,6144,0.008969777988062965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,10240,0.014408000641398959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,5120,0.0105262224872907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,65536,0.14861244625515407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,10240,0.01640888883007897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,64,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,4096,0.009346666435400644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,2560,0.03462222218513489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,4096,0.012886222038004132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,16384,0.01829866733815935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,1536,0.014504889647165934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,6144,0.012828444441159567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,12288,0.03308622373474969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,8192,0.05744889047410753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,512,0.011391110718250275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,7168,0.03418666786617703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,4096,0.011429333024554782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,2560,0.007466666400432587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,8192,0.03663644525739882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,768,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,7168,0.02345866627163357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,65536,0.19443910651736787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,3072,0.0112631113992797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,12288,0.034859554635153875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,128,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,8192,0.018060444129837882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,12288,0.060200889905293785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,6144,0.021871111459202234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,512,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,512,0.0028231110837724474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,2560,0.06224799818462796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,2048,0.006280888699822956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,6144,0.13569600052303737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,65536,0.057913780212402344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,768,0.0025333333760499954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,128,0.0148062225845125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,128,0.0021199999998013177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,2048,0.014462222655614218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,65536,0.09535466962390476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,32,0.012051555845472546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,65536,0.17629599571228027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,2048,0.014087999860445658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,7168,0.010012444522645738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,6144,0.011083555718262991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,3584,0.008013333711359236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,1024,0.013233777549531726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,3584,0.007353777686754863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,8192,0.02402399977048238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,6144,0.025402666793929204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,16384,0.02026844521363576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,768,0.00664533343580034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,3072,0.10256622234980266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,512,0.007287110719415877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,2048,0.019181332654423185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,5120,0.009748444788985783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,256,0.016882666283183627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,16384,0.6878479851616753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,12288,0.02898933490117391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,4096,0.08881511290868123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,65536,0.07904355393515693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,3584,0.007439110842016008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,2560,0.006999111009968652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,6144,0.01480888823668162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,10240,0.1577964491314358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,3584,0.02867022156715393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,10240,0.8530426555209689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,2560,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,64,0.0017955555684036678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,3584,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,1536,0.013086222112178802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,8192,0.009318222602208456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,1024,0.005944889038801193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,2560,0.007947555846638149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,8192,0.011038222246699862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,512,0.006727111008432176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,3072,0.02134311033619775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,3584,0.007344000041484833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,128,0.014169777433077494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,10240,0.005929777605666055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,1024,0.0031840000301599503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,128,0.014148443937301636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,7168,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,8192,0.022154667311244543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,512,0.0026088888860411118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,256,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,3072,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,12288,0.007263999846246507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,256,0.01682222220632765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,12288,0.04039377636379666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,64,0.0019262221952279408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,6144,0.009000000026490953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,768,0.00833511104186376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,16384,0.05362488826115926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,128,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,256,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,16384,0.04520977867974175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,512,0.0028275555620590844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,2048,0.06614844666586982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,512,0.014142221874660917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,65536,0.12452533509996201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,8192,0.02367555598417918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,4096,0.016404444972674053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,8192,0.016895999511082966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,128,0.011743110915025076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,1536,0.032136887311935425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,1536,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,3584,0.09085333347320557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,65536,0.1969431108898587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,2048,0.03909066650602553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,4096,0.04667733444107903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,3072,0.013831111292044321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,512,0.004919111314747068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,2560,0.016812443733215332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,2560,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,256,0.005938666562239329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,16384,0.028804444604449805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,16384,0.0432533323764801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,6144,0.011521778172916837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,2560,0.029755555921130713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,8192,0.0125928885406918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,1536,0.02572622232966953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,3072,0.007599111232492659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,16384,0.03654399845335219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,128,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,10240,0.0120328888297081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,2560,0.013440888788965015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,2560,0.015672889020707872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,3584,0.007272889216740926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,1024,0.007247111035717859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,768,0.0058755555914507965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,3584,0.02614044480853611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,10240,0.061475555102030434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,2560,0.00903200027015474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,7168,0.018869333797030978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,768,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,10240,0.11065422164069282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,16384,0.119558228386773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,768,0.00666844430896971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,512,0.019574221637513902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,1536,0.005954666684071223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,7168,0.013266666895813413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,16384,0.022407111194398668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,4096,0.00797333319981893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,3072,0.01707555519209968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,1536,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,65536,0.09628888633516099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,6144,0.014145778285132514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,65536,0.388791110780504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,12288,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,10240,0.045463999112447105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,1024,0.017906667457686532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,768,0.005714666512277391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,10240,0.06295555830001831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,64,0.0022151112142536375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,3072,0.007341333561473423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,768,0.013192888763215808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,1024,0.04299911194377475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,5120,0.00847999999920527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,768,0.0028862222615215513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,5120,0.009146666361225976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,2048,0.006534222099516127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,8192,0.06149333053165012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,12288,0.05188799897829691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,5120,0.11129155423906113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,512,0.0053013331360287136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,5120,0.04122755593723721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,2560,0.010679999987284342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,32,0.006261333409282897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,5120,0.014145778285132514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,768,0.011017777853541903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,8192,0.011078221930397881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,7168,0.0317813323603736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,12288,0.8628480169508191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,3584,0.026808000273174707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,6144,0.029097778929604426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,2048,0.030605332718955144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,8192,0.08304266797171699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,10240,0.012415111064910889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,3584,0.010480889015727572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,64,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,256,0.005149333427349727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,5120,0.012679110798570844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,1024,0.015828443898095023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,2560,0.007130666739410824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,3072,0.008182221816645728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,768,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,32,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,8192,0.01824977828396691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,1536,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,4096,0.00796888851457172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,256,0.0022008888837363985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,768,0.004017777740955353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,2560,0.06784177488750882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,2560,0.007033777733643849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,65536,0.08115377691056994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,64,0.0038399998512532976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,10240,0.02533600065443251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,128,0.014492443866199918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,256,0.0022622222701708474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,768,0.005379555539952383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,16384,0.014708444476127625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,3072,0.08082489172617595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,512,0.011414222419261932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,6144,0.011702222128709158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,768,0.0028773333049482773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,2560,0.02958044409751892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,1024,0.012271111210187277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,4096,0.03169688913557265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,3584,0.016120000018013846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,128,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,12288,0.02847822176085578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,512,0.0025155554629034465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,12288,0.023021333747439917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,1024,0.1596213314268324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,7168,0.015836444165971544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,10240,0.19189955128563774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,5120,0.005618666609128316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,5120,0.05394577648904589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,10240,0.018938667244381376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,7168,0.017880888448821176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,3584,0.03011466728316413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,16384,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,3584,0.020223110914230347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,8192,0.11898399723900689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,256,0.004935111022657818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,8192,0.04946310983763801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,10240,0.15967110792795816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,768,0.03626577721701728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,768,0.0032044444233179092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,128,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,12288,0.02325511144267188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,4096,0.10544088813993667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,64,0.005050666630268097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,768,0.0027946666296985415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,1024,0.020419556233617995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,1024,0.23041065533955893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,7168,0.007256888680987888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,1024,0.040543999936845564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,2560,0.008640000389681922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,64,0.014098667436175875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,2048,0.007304000357786815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,8192,0.025120000044504803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,6144,0.04935822221967909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,2560,0.02257599929968516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,10240,0.012067555553383298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,5120,0.00851999968290329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,3072,0.007269333634111617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,7168,0.014427555931939019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,5120,0.020624889267815482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,3584,0.03052088949415419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,512,0.006245333287451003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,8192,0.013752000199423896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,65536,0.08176177740097046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,1024,0.0035048888789282907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,16384,0.010710222025712332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,8192,0.01583377851380242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,12288,0.22928977012634277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,12288,0.10443911287519668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,7168,0.01734133395883772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,2048,0.009383111364311641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,10240,0.026151110728581745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,12288,0.010679999987284342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,256,0.007511999871995714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,2048,0.07697511381573148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,3584,0.11481244034237331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,2048,0.019102222389645047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,12288,0.019125334090656705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,512,0.00976533359951443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,10240,0.018956444329685636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,4096,0.01962488889694214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,7168,0.014919999572965833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,7168,0.02405066622628106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,6144,0.009357333183288574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,2048,0.010023111270533668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,10240,0.02015911042690277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,1536,0.007152888509962294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,12288,0.013990221752060784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,65536,0.2149288919236925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,128,0.009071110851234859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,512,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,7168,0.01034311122364468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,65536,0.14943110942840576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,7168,0.02232444451914893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,1024,0.00629688882165485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,128,0.005621333503060871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,768,0.01477688882086012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,4096,0.0318195554945204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,3072,0.005937777873542573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,2048,0.01864622202184465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,6144,0.017095999585257638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,768,0.014836443795098199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,1024,0.005912888795137405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,512,0.005672889037264719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,3072,0.02438844409253862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,3072,0.021967111362351313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,128,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,768,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,1536,0.006160000132189856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,3072,0.0069644442862934535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,4096,0.00795911086930169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,768,0.029281778468026057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,12288,0.0192586663696501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,32,0.002475555572244856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,12288,0.017886221408843994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,32,0.0018159999615616268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,65536,0.09619822104771932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,7168,0.02160355614291297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,768,0.007312888900438945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,65536,0.018844443890783522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,3584,0.00500266667869356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,2560,0.09412888685862224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,2048,0.02682488825586107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,1536,0.010746666954623329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,16384,0.05283288823233711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,1024,0.006213333457708359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,12288,0.02436888880199856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,1024,0.0069875551594628235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,3072,0.005304000029961268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,128,0.005882666756709416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,128,0.014162666267818876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,1024,0.004545777622196409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,128,0.014516444669829475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,16384,0.4756124284532335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,1536,0.006009777800904379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,4096,0.009992000129487779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,12288,0.01586844523747762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,512,0.005205333232879639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,512,0.023950222465727065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,4096,0.007999999655617608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,128,0.014161777165200977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,65536,0.09191377957661946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,65536,0.05336711141798231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,128,0.00554044461912579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,3072,0.006994666324721442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,12288,0.012416889270146688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,6144,0.00508711114525795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,768,0.01757777730623881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,16384,0.10833955473370022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,6144,0.017687999539905123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,6144,0.005731555736727185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,64,0.004919111314747068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,12288,0.04009333252906799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,16384,0.006918222126033571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,7168,0.010054222411579555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,6144,0.011014222270912595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,512,0.0052586665583981415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,7168,0.005296000176005893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,12288,0.21536710527208117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,7168,0.41976533995734316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,16384,0.028502222564485338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,768,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,128,0.10086666875415379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,128,0.0018008889423476325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,7168,0.010013333625263637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,512,0.022203556365436975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,3072,0.03650577863057455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,256,0.002208000048995018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,4096,0.013815999858909182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,10240,0.032251556714375816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,5120,0.05477866530418396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,65536,0.08256266514460246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,12288,0.03881066706445482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,1536,0.03621244430541992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,2560,0.00655377780397733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,16384,0.058430221345689565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,16384,0.24716001086764863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,12288,0.05084088775846693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,256,0.014136888914638095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,768,0.01899111105336083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,1024,0.01581333412064446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,256,0.002616000051299731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,8192,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,8192,0.013832888669437833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,16384,0.024460444847742718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,3584,0.007638221813572778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,4096,0.010788444843557147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,10240,0.007341333561473423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,10240,0.019317333896954853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,3584,0.01591466698381636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,1024,0.0063662222690052455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,12288,0.007308444215191736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,768,0.01723377737734053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,10240,0.04584355486763848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,3072,0.019946666227446664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,2560,0.10973244243197972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,128,0.0052728888889153796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,8192,0.031195554468366835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,1024,0.05758755736880832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,6144,0.010112889111042023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,7168,0.042472889026006065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,256,0.012732444538010491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,65536,0.29029689894782174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,256,0.01681688924630483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,2560,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,2560,0.02029511166943444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,256,0.0024133332901530797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,128,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,512,0.014885332849290637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,1536,0.0038764443662431505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,5120,0.009014222357008193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,1024,0.004864888886610667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,6144,0.019871999820073444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,2560,0.008719999757077958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,3584,0.01718488832314809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,32,0.0028151110228565005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,8192,0.019259555472267997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,3072,0.008996444443861643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,512,0.011763555308183035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,1536,0.004284444368547863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,768,0.026702221896913316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,6144,0.04243466589185926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,768,0.005886222339338726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,512,0.006238222122192383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,2560,0.02165688905451033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,65536,0.06547288762198554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,12288,0.010953777366214328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,1024,0.006055111272467508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,1024,0.004574222283230888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,256,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,7168,0.02936089038848877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,3584,0.009416888985368941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,5120,0.0372559991147783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,3584,0.10057066546546088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,64,0.0018399999373488957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,1024,0.003232888877391815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,1536,0.0034888887570963967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,2048,0.007873777714040544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,2560,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,1536,0.01735466718673706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,2560,0.00665244460105896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,1536,0.009024000002278222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,2560,0.05613955524232653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,768,0.02176800039079454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,64,0.009644444617960189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,1024,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,10240,0.05370311273468865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,2560,0.022121777137120564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,6144,0.04179999894566006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,1536,0.00869599978129069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,4096,0.17170222600301108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,2048,0.007608888877762689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,2560,0.010074666804737514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,1024,0.01833866702185737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,768,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,7168,0.02840355700916714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,7168,0.009656888743241629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,3584,0.007375111182530721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,2048,0.006605333338181178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,3072,0.02018044392267863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,2048,0.007739555504586961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,512,0.009323555562231276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,4096,0.01680888897842831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,64,0.0017991111510329777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,5120,0.011000889043013254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,3072,0.013140444126394061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,8192,0.018717333674430847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,256,0.004869333157936732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,2048,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,16384,0.023327999644809302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,12288,0.14643643962012395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,3072,0.010441777606805166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,7168,0.00999466660949919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,1024,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,128,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,16384,0.09564710987938775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,8192,0.011725333001878528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,6144,0.010667555862002902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,128,0.014095111025704278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,5120,0.013269333375824822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,3584,0.036534223291609026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,768,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,128,0.02720888952414195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,256,0.022700443863868713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,16384,0.01571111049917009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,1024,0.005919999960396025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,5120,0.008987555901209513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,512,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,8192,0.016843555702103507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,4096,0.02000711030430264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,64,0.0021217777911159727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,3584,0.035537779331207275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,128,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,65536,0.018869333797030978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,64,0.001968888979819086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,2560,0.03638933433426751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,7168,0.01275199982855055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,6144,0.0270480000310474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,12288,0.039579557047949895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,12288,0.005469333380460739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,768,0.01715377800994449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,32,0.001818666648533609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,128,0.014519999424616495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,2048,0.01440711153878106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,10240,0.016179554992251925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,4096,0.008681777450773451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,1536,0.0035537779331207275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,256,0.012431110772821637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,5120,0.016136000553766888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,1536,0.013051555388503604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,256,0.005895111295912001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,12288,0.013541333377361298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,256,0.00906133320596483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,10240,0.011350221931934357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,6144,0.010010666317409938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,1536,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,65536,0.1514942248662313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,10240,0.013218666944238873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,65536,0.012333333492279053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,2560,0.006613333192136552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,3072,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,128,0.009024000002278222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,6144,0.010354667074150508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,128,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,512,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,2048,0.006103111224042044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,12288,0.028240889310836792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,512,0.009540444446934594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,8192,0.019280888968043856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,2048,0.0730088882976108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,16384,0.05156622330347697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,12288,0.01478666729397244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,8192,0.015173332558737861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,6144,0.03348799877696567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,12288,0.06185866726769341
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,64,0.0018639999131361644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,64,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,12288,0.022663111488024395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,4096,0.007660444411966536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,256,0.0029333333174387612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,5120,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,16384,0.02331200076474084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,512,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,3584,0.11519466506110297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,2560,0.01687466601530711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,768,0.010033778018421596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,5120,0.014088888963063559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,64,0.006591111007663939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,2560,0.016374222106403775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,4096,0.02146666745344798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,64,0.002471111093958219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,7168,0.008645333349704742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,4096,0.015469332536061605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,10240,0.03151999910672506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,1024,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,65536,1.1568373574150932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,8192,0.012601777911186218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,128,0.006607999818192587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,1536,0.008360889222886827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,768,0.03298666742112901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,256,0.015187554889255099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,7168,0.010037333601050908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,65536,0.2482933203379313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,10240,0.015160888433456421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,5120,0.008786666724416945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,2560,0.009995555712117089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,64,0.005623111294375525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,3584,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,128,0.005769777629110549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,32,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,2560,0.006223111102978389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,4096,0.015508444772826301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,5120,0.013901333014170328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,4096,0.005544888890451855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,4096,0.2769715521070692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,65536,0.12185777558220758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,2560,0.007100444700982835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,6144,0.008989333278603023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,16384,0.024174221687846716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,256,0.019288000133302476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,3584,0.014282666974597506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,2560,0.006256000035338932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,2048,0.007296000089910295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,256,0.016840888394249808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,16384,0.0366906656159295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,3072,0.034832000732421875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,65536,0.4511920081244574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,512,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,8192,0.010547555155224271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,3584,0.04093066520161099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,3584,0.009360888765917884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,1536,0.04111466805140177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,1536,0.012282667060693106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,7168,0.03610666592915853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,6144,0.016567111015319824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,3072,0.09024622042973836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,4096,0.00790666706032223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,2048,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,65536,0.1799431112077501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,512,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,3072,0.015113777584499784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,3072,0.004579555657174853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,12288,0.01312711089849472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,2560,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,512,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,1536,0.0038613333470291565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,768,0.015471999843915304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,8192,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,64,0.0058755555914507965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,32,0.0038604442444112566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,65536,0.07971288760503133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,4096,0.006621333460013072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,65536,0.15760444270239934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,5120,0.026421333352724712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,65536,0.06523378027809991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,1536,0.012844444149070315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,65536,0.4734346601698134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,12288,0.11931467056274414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,32,0.003181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,65536,0.05669689178466797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,768,0.005601777798599667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,768,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,3584,0.023303111394246418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,16384,0.03663377629386054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,1536,0.061101330651177295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,768,0.005959110955397288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,2560,0.031812445984946355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,512,0.011403555671374003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,7168,0.0818915565808614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,5120,0.01015733347998725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,256,0.011687111523416309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,1536,0.0038968887594011095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,1024,0.0151555554734336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,16384,0.024996444582939148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,8192,0.032446222172843084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,16384,0.02572888963752323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,2560,0.0042035554846127825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,4096,0.012723555167516073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,65536,0.20443911022610137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,7168,0.16470489237043592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,512,0.014208889669842191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,32,0.0052800000541739995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,1536,0.015792888071801927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,65536,0.5401688681708442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,2048,0.011115555961926779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,7168,0.02648711038960351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,2048,0.022299556268586054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,3072,0.00795911086930169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,7168,0.014820444915029736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,12288,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,16384,0.03371466530693902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,12288,0.02847555610868666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,12288,0.24532977739969888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,128,0.006270222365856171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,8192,0.007981333467695449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,512,0.032391998502943255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,12288,0.02058844433890449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,3072,0.011002666420406766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,8192,0.01723022262255351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,12288,0.07924266656239827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,1024,0.005963555640644497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,3072,0.007632888853549957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,4096,0.015683554940753512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,1024,0.018907555275493197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,12288,0.03570222192340427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,5120,0.009002666506502364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,1536,0.005922666854328579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,65536,0.12799378236134848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,768,0.004554666578769684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,1536,0.019575110740131803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,768,0.006219555520349079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,32,0.0021688888470331826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,128,0.00309955556359556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,5120,0.013109332985348172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,2048,0.014397333065668741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,6144,0.03017510970433553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,768,0.028392000330819026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,512,0.002479111154874166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,6144,0.009359111388524374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,256,0.014479110638300577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,8192,0.024246222443050806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,64,0.0057911111248864075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,32,0.005382222019963794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,1024,0.009498666558000777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,64,0.0038853333228164246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,16384,0.020603555772039626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,3072,0.006986666884687211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,256,0.014307555225160388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,5120,0.008108444511890411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,5120,0.008336889247099558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,8192,0.014831110835075378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,16384,0.028191109498341877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,1024,0.0038977778620190094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,512,0.012778666284349231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,2560,0.006635555376609166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,1536,0.006254222244024277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,5120,0.009835555321640438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,65536,0.16189244058397081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,2048,0.018272888329293992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,64,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,10240,0.018195556269751657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,768,0.01888088881969452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,3072,0.004272888931963178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,1536,0.019419555862744648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,4096,0.00830577810605367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,12288,0.46347909503512913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,64,0.008971555365456475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,3072,0.018178666631380718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,7168,0.01125600023402108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,128,0.01718577742576599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,65536,0.07762133412890963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,12288,0.015085332923465304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,2048,0.0239955551094479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,12288,0.029503110382292006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,256,0.004936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,1024,0.004763555609517627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,2048,0.00664533343580034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,7168,0.3462053404914008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,512,0.011397333608733283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,10240,0.011732444167137146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,6144,0.06642755534913805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,768,0.004567999806669024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,8192,0.04011911153793335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,3072,0.017996443642510306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,5120,0.005543111099137201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,2560,0.0069751110341813825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,1024,0.005583999885453119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,2048,0.0042408886882993905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,256,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,128,0.006615110983451207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,768,0.024439111351966858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,2048,0.004553777890072929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,4096,0.007652444144090016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,64,0.002473777780930201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,5120,0.010264000131024255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,512,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,8192,0.012759110993809171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,7168,0.009663110805882348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,2048,0.013425778183672162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,3072,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,2048,0.006299555715587404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,32,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,16384,0.03288977675967746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,16384,0.022670222653283015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,32,0.005252444495757421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,7168,0.016984888248973422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,16384,0.023752000596788194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,3584,0.02810133496920268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,65536,0.07416177458233304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,2560,0.019239111079110038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,3584,0.03345244460635715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,12288,0.02104622291194068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,7168,0.010642666783597736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,512,0.04598044355710348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,12288,0.11919111675686306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,2048,0.0299893319606781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,4096,0.00795555528667238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,3072,0.01891288823551602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,10240,0.12425422668457031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,7168,0.030015110969543457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,2048,0.021003555920388963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,1024,0.006215999937719769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,10240,0.015013333823945789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,6144,0.02167555524243249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,3584,0.01924000018172794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,10240,0.050739556550979614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,7168,0.014081777797804939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,7168,0.027608889672491286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,7168,0.0762959983613756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,3584,0.019560888409614563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,3584,0.016920000314712524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,1024,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,768,0.016197333733240765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,1024,0.006203555398517185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,2048,0.02420355545149909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,512,0.0028773333049482773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,1024,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,32,0.006588444527652528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,3072,0.057906667391459145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,2048,0.014520888527234396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,8192,0.037053333388434515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,2560,0.01925155520439148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,2048,0.005616000129116907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,2560,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,8192,0.04026666614744399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,12288,0.01497866710027059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,2048,0.004563555535342959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,8192,0.01827822294500139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,64,0.002512000087234709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,2048,0.0120000003112687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,12288,0.05362844467163086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,512,0.004952888935804367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,3072,0.013923555612564087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,65536,0.10333422157499526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,32,0.002536888958679305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,256,0.006413333531883027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,12288,0.027836445305082533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,8192,0.04638311266899109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,4096,0.41590311792161727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,5120,0.029149333635965984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,4096,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,65536,0.054808888170454234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,3584,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,768,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,3584,0.011418666276666852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,1024,0.01382844481203291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,128,0.013462222284740873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,4096,0.034359110726250544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,1536,0.013501333693663279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,4096,0.014764444695578681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,12288,0.013722666435771518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,4096,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,16384,0.10716622405582005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,6144,0.5039164225260416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,768,0.00794577764140235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,1536,0.012223111258612739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,10240,0.019661333825853135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,2048,0.00830577810605367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,768,0.017631111873520743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,5120,0.1398435566160414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,8192,0.01014755583471722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,3584,0.024098666177855596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,1024,0.016778666112158034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,256,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,5120,0.01239377756913503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,1536,0.007248000138335758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,5120,0.00835111074977451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,10240,0.035960889524883695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,8192,0.12409599622090657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,4096,0.009317333499590555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,128,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,768,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,10240,0.011368888947698804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,10240,0.3715493414137099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,2560,0.005624889085690181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,6144,0.01681688924630483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,3584,0.026036444637510512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,65536,0.11154845025804307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,4096,0.013011555704805585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,8192,0.04008888867166307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,2560,0.008395555946562026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,128,0.013737777868906656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,3584,0.012062222593360476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,3072,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,64,0.006266666783226862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,1536,0.012376888758606382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,128,0.014168888330459595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,16384,0.025383111503389146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,8192,0.048828442891438804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,65536,0.046204444434907704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,3584,0.010006222460005019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,6144,0.029993779129452173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,6144,0.009022222624884712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,1024,0.00904888908068339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,2560,0.01962844365172916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,65536,0.37393866644965273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,768,0.0037271111375755737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,3072,0.008459555606047312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,7168,0.6805075539482964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,256,0.0028204443968004654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,7168,0.010489777558379702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,65536,0.26319289207458496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,4096,0.012296888563368054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,1536,0.006945778098371293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,65536,0.10636978016959296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,7168,0.0229751103454166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,512,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,7168,0.00999022192425198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,1536,0.013248888982666863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,8192,0.00814222213294771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,12288,0.013887999786270989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,64,0.0026311110705137253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,768,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,65536,0.07558400101131864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,4096,0.009702222214804756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,512,0.004957333207130432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,4096,0.025154666768179998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,1536,0.05088622371355692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,256,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,5120,0.014950222439236112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,1024,0.012712889247470431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,10240,0.017133333616786532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,512,0.0025946667624844448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,768,0.01721333298418257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,64,0.013080889152155982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,512,0.013433777623706393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,3072,0.007243555453088548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,2048,0.006604444649484422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,4096,0.09151111046473186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,6144,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,768,0.0029155556112527847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,2048,0.007041778001520369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,6144,0.09596266349156697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,3072,0.007659555309348636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,1024,0.008397333323955536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,128,0.02161600026819441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,16384,0.045957333511776395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,3072,0.03124533428086175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,8192,0.043805334303114146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,6144,0.01834133267402649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,12288,0.01588355501492818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,2048,0.009402666654851701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,16384,0.2782222165001763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,512,0.030199110507965088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,768,0.007619555625650618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,1536,0.008722667064931657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,7168,0.02145511077509986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,1536,0.007184888753626082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,3584,0.00794311116139094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,768,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,2560,0.013750221994188098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,32,0.0019600000232458115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,12288,0.06704444355434842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,256,0.002483555633160803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,2560,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,4096,0.008017777568764156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,5120,0.10728266504075791
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,1024,0.008203555312421586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,2048,0.009604444106419882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,768,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,12288,0.013774221969975365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,512,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,3584,0.03760711020893521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,768,0.017662222186724346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,3584,0.005616000129116907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,5120,0.01314399970902337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,1024,0.007621333003044128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,2048,0.004540444248252445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,32,0.004920000003443824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,6144,0.03713599840799967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,6144,0.012424888710180918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,2048,0.01818933255142636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,4096,0.01586666703224182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,768,0.016427555018001132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,256,0.012373333175977072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,10240,0.03966844413015578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,3584,0.011239999698268043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,768,0.003020444470975134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,4096,0.028564443190892536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,10240,0.038592000802357994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,1024,0.05845955345365736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,1024,0.0177875558535258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,4096,0.007679111427730984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,8192,0.24681332376268175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,64,0.0025395556456512874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,2048,0.008040000167157915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,12288,0.02980888883272807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,7168,0.007925333248244392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,512,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,65536,0.02640977833006117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,128,0.007975999679830339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,16384,0.19126666916741264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,5120,0.030621333254708186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,768,0.013295110729005603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,8192,0.143394668896993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,65536,0.09062933259540135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,256,0.0063564446237352155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,1536,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,256,0.02069244450993008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,5120,0.020437333318922255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,256,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,10240,0.020250666472646926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,8192,0.01716711123784383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,5120,0.06379110945595635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,1536,0.01274222218328052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,3584,0.07552622424231635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,1024,0.005913777897755305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,4096,0.023003555006451074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,7168,0.014714666538768344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,768,0.0032088889016045462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,6144,0.016545777519543965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,5120,0.020346666375796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,10240,0.024004444479942322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,512,0.011014222270912595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,768,0.019483556350072224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,2048,0.004228444563017951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,5120,0.016158221496476066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,16384,0.0273617770936754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,4096,0.030798223283555772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,5120,0.004561777744028303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,4096,0.05244177911016676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,2560,0.00775733341773351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,16384,0.024805333879258897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,2048,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,10240,0.01181777815024058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,2560,0.006592000110281839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,1536,0.027535999814669292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,16384,0.09470044242011176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,2560,0.1906693379084269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,65536,0.08801244364844428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,7168,0.02773422168360816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,8192,0.054634667105144925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,64,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,1024,0.005583999885453119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,768,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,128,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,1024,0.011429333024554782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,2048,0.04311555624008179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,1024,0.01635199950801002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,7168,0.017853332890404593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,768,0.005952000204059813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,3584,0.01796000036928389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,1024,0.0032035555276605818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,7168,0.029521776570214167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,3072,0.010667555862002902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,1024,0.015153777268197803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,256,0.004952888935804367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,768,0.01689511040846507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,12288,0.01885955532391866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,8192,0.0058746664888328975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,1536,0.07554488711886935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,65536,0.09047644668155247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,128,0.013809777796268463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,6144,0.015844444433848064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,2048,0.004540444248252445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,12288,0.012740444805887012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,3584,0.007647111184067196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,64,0.008479110896587372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,65536,0.05681155787573921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,8192,0.017904000149832833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,768,0.01573511130279965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,128,0.0020995556066433587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,8192,0.045913779073291354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,512,0.013076444466908773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,5120,0.0063235556913746735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,65536,0.4851004282633464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,768,0.01720177796151903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,1024,0.006288888967699475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,128,0.015770667129092746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,6144,0.016494222813182406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,1536,0.006592888798978593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,64,0.0038657776183552216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,32,0.0058622223635514574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,1024,0.016467556357383728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,7168,0.013162666724787818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,16384,0.031143112315071955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,3584,0.061639110247294106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,1024,0.025627555118666753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,65536,0.020263999700546265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,65536,0.098399109310574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,3584,0.014697778556081982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,256,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,768,0.006005333529578314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,256,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,6144,0.006583111153708563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,2560,0.008279110822412703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,256,0.006590222318967183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,64,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,512,0.011063999599880643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,256,0.00519466648499171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,768,0.005576000031497743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,3584,0.0075804442167282104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,65536,0.009300444689061906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,32,0.00795377790927887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,128,0.020672000116772123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,2048,0.02328799996111128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,6144,0.06707466973198785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,2048,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,1536,0.015122666954994202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,256,0.0029111111329661477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,1024,0.012142221960756512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,65536,0.05500622259245979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,7168,0.009653333160612319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,7168,0.005542222410440445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,32,0.002546666603949335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,4096,0.009469333622190688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,1536,0.013071999781661563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,4096,0.015815999772813585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,2560,0.006613333192136552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,65536,0.14994666311475965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,5120,0.02437866727511088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,512,0.05500444438722399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,16384,0.1706951061884562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,2560,0.014404444230927361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,12288,0.018946665856573317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,768,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,1536,0.0038888889054457345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,768,0.0028817777832349143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,16384,1.3077164755927193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,1536,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,256,0.013420444395807056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,1536,0.018237334158685472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,4096,0.008667555948098501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,12288,0.03388533327314589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,4096,0.13298488987816706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,2560,0.015663110547595553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,256,0.00924800005224016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,1024,0.0031964443624019623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,10240,0.01759999990463257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,3584,0.007800000409285228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,256,0.005057777795526716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,12288,0.02420711186197069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,128,0.014742222097184924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,1536,0.006263999889294307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,512,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,12288,0.4813946617974176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,6144,0.01613066593805949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,6144,0.010695999695195092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,2560,0.011379555695586734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,2560,0.03537066777547201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,256,0.020602666669421725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,256,0.01687733332316081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,768,0.017530666457282174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,5120,0.009704888694816166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,3072,0.022092445029152766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,64,0.002064888882968161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,3584,0.040384001202053495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,3072,0.014558222558763293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,16384,0.37720444467332626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,256,0.004940444396601783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,128,0.007248889240953658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,768,0.006132444573773279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,1024,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,6144,0.013718222578366598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,4096,0.006605333338181178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,768,0.009311111436949836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,128,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,2048,0.004585777719815572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,12288,0.009397333694828881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,65536,0.00656088896923595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,3584,0.0076044441925154785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,4096,0.0125591109196345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,12288,0.019556444552209642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,2560,0.030947556098302204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,128,0.005922666854328579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,3072,0.025095999240875244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,2560,0.05307999915546841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,128,0.014780445231331719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,10240,0.2822311189439562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,6144,0.013921777407328287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,16384,0.023697778582572937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,12288,0.014106666048367819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,512,0.012154666913880242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,768,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,3072,0.01071200023094813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,16384,0.021140444609853957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,128,0.007295110987292395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,5120,0.12190843953026666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,256,0.03793066740036011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,10240,0.012016889121797351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,16384,0.1532720062467787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,1024,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,32,0.0021946666141351065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,6144,0.010300444232092964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,4096,0.00794311116139094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,3584,0.007981333467695449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,768,0.016125332978036668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,256,0.005628444254398346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,4096,0.00795111142926746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,512,0.018644443816608854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,3072,0.007328888608349695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,16384,0.03152889013290405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,256,0.024467556013001338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,5120,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,512,0.005580444302823808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,8192,0.020271999968422782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,64,0.005299555758635203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,1024,0.011759999725553723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,32,0.005888000130653381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,3584,0.011062222222487131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,5120,0.037696888049443565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,7168,0.03803911142879062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,7168,0.022967999180157978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,2048,0.014883556299739413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,512,0.005249777601824866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,16384,0.03535822365019057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,3584,0.012462221913867526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,768,0.015971556305885315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,64,0.001971555459830496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,6144,0.03966222206751505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,512,0.002612444468670421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,256,0.005576000031497743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,4096,0.014769777655601501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,10240,0.013287111288971372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,512,0.009820444716347588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,6144,0.009715555442704095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,32,0.015843555331230164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,7168,0.13712178336249456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,6144,0.015230221880806817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,32,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,7168,0.030524445904625788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,10240,0.04227466715706719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,8192,0.030493332280053034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,3072,0.022983110613293115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,256,0.024903999434577093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,32,0.006568000134494569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,8192,0.013178666432698568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,65536,0.11921600500742595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,2048,0.013459555804729462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,8192,0.0042408886882993905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,1536,0.027238221632109746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,64,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,512,0.005790222022268508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,3584,0.007610666255156199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,4096,0.011027555498811932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,12288,0.014097778333557976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,512,0.01237155579858356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,10240,0.02477955487039354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,2048,0.03877244393030802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,64,0.006224888894293044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,7168,0.01682044400109185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,1536,0.006255110932721033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,768,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,2560,0.015802666544914246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,65536,1.5181759728325739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,512,0.011019555230935415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,16384,0.030307554536395605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,32,0.016153777639071148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,3072,0.07078222433725993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,768,0.018578666779730055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,1536,0.05558844407399496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,10240,0.03174933460023668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,256,0.004868444469239977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,8192,0.166101336479187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,8192,0.010220444036854638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,5120,0.01314933349688848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,10240,0.02072355482313368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,2560,0.008714666797055138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,128,0.001822222231162919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,7168,0.004855111241340637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,4096,0.30079110463460285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,7168,0.01755555636352963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,3072,0.007605333295133378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,3072,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,16384,0.06360089116626315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,8192,0.04909866717126635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,4096,0.007607111500369177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,12288,0.011309333145618439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,128,0.014454222387737699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,32,0.003154666680428717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,768,0.0042364444169733255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,128,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,8192,0.0355422231886122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,4096,0.014768888552983602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,7168,0.010085333552625444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,12288,0.025055110454559326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,3584,0.011768889096048145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,65536,0.09114755524529351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,4096,0.012886222038004132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,6144,0.03832177652253045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,2048,0.00738844441043006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,16384,0.08229244417614408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,16384,0.10320711135864258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,768,0.01310222182008955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,2560,0.02439911166826884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,65536,0.07505866554048327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,2560,0.06450221935908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,10240,0.05124355687035454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,1024,0.04103199972046746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,16384,0.07657866345511542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,1024,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,1536,0.004064889003833135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,768,0.00794488853878445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,768,0.003152888889114062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,256,0.011725333001878528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,256,0.004579555657174853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,10240,0.020240889655219186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,768,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,4096,0.026055110825432673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,12288,0.01645511057641771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,64,0.0017991111510329777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,4096,0.03453955716556973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,12288,0.169868442747328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,3584,0.013800000150998434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,128,0.005192888693677055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,10240,0.011716444459226398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,128,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,3584,0.011751111182901593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,2560,0.016921778519948322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,8192,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,256,0.005615111026499007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,2048,0.0069413334131240845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,5120,0.015808888607554965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,6144,0.009691555466916826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,10240,0.00630044440428416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,256,0.0028764444092909494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,128,0.005965333431959152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,2560,0.008113777471913232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,7168,0.04050755500793457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,2048,0.01888533267709944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,64,0.0058240000572469496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,8192,0.023835556374655828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,8192,0.011991110940774282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,1024,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,1536,0.012868444124857584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,7168,0.030200888713200886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,8192,0.0295688874191708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,512,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,1536,0.018223111828168232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,1024,0.01161688897344801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,3072,0.008004444340864817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,12288,0.025042666329277888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,3072,0.022337777747048274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,6144,0.00904444439543618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,65536,0.016498666670587327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,6144,0.02055911057525211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,128,0.005647111270162795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,256,0.005948444621430502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,8192,0.08906577693091498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,1536,0.017176000608338248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,256,0.01682044400109185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,32,0.00656088896923595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,2048,0.006228444476922353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,1536,0.006396444307433234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,512,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,768,0.01550222271018558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,3072,0.007667555577225155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,4096,0.027811557054519653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,12288,0.028560890091790095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,6144,0.010975999964608086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,768,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,128,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,256,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,3584,0.03221599923239814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,65536,0.11903999911414252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,10240,0.03231733375125461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,3584,0.00795733349190818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,12288,0.16495733790927464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,128,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,4096,0.02986933456526862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,5120,1024,0.020633776982625324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,6144,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,3584,0.5057226816813151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,4096,0.030518223841985066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,16384,0.2312862343258328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,3584,0.010015111002657149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,8192,0.011043555206722684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,128,0.007284444239404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,1536,0.01309333327743742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,768,0.05370577838685778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,256,0.014833778142929077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,3072,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,65536,0.5350515577528212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,2048,0.004263111286693149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,256,0.016180444094869826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,8192,0.006616888774765863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,12288,0.016063110695944894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,12288,0.03898222247759501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,512,0.007317333585686154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,32,0.0018159999615616268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,3072,0.01642666591538323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,65536,0.13220000267028809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,1536,0.004613333278232151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,7168,0.023944889505704243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,8192,0.039112889104419284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,65536,1.1358790927463107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,5120,0.01238844460911221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,1024,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,65536,0.07950133085250854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,32,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,1024,0.019488000207477145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,12288,0.02197244432237413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,1024,0.012435555458068848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,768,0.023656888140572443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,12288,0.01955911186006334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,7168,0.3529093265533447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,256,0.002528000002106031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,3072,0.00798044436507755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,16384,0.02385155525472429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,10240,0.16366044680277506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,768,0.013080889152155982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,1536,0.015458666616015963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,768,0.024504888388845656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,10240,0.016340444485346477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,5120,0.011941333611806234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,65536,0.21663999557495117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,2048,0.010701333483060202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,3584,0.05353066656324598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,7168,0.011734222372372946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,8192,0.10808177789052327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,3584,0.014319999350441827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,65536,3584,0.13062666522132024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,4096,0.015082667271296183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,768,0.049032890134387545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,256,0.013450666434235044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,512,0.025478222303920325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,65536,0.15180177158779568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,3072,0.007347555624114142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,16384,0.033597333563698664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,5120,0.027505778604083594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,12288,0.01683022247420417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,8192,0.04447999927732679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,64,0.00738666703303655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,5120,0.021894221504529316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,768,0.01754577789041731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,2048,0.1328559981452094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,4096,0.024894222617149353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,768,0.006628444625271692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,32,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,8192,0.011746666497654386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,2048,0.022449778185950384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,512,0.021606221795082092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,8192,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,512,0.002571555475393931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,256,0.004545777622196409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,12288,0.020986666282018025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,10240,0.014119999276267158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,16384,0.09975733359654744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,10240,0.19555555449591744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,32,0.007605333295133378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,2560,0.01856355534659492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,12288,0.04344266653060913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,1536,0.01143555591503779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,768,0.017521777086787753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,3584,0.014821334017647637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,10240,0.49176353878445095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,768,0.013427555561065674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,3072,0.017859554953045316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,1536,0.004594666676388847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,16384,0.04904266529613071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,6144,0.0403439998626709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,1536,0.005966222120655908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,768,0.005990222096443176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,128,0.014115555418862237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,32,0.005569777968857024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,5120,0.017138666576809354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,3072,0.013348444468445249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,6144,0.020390222469965618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,256,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,16384,0.031160000297758315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,2048,0.00865955568022198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,8192,0.01648622254530589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,128,0.004886222382386525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,7168,0.021751999855041504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,128,0.004944889081848992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,512,0.0099973330895106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,256,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,2560,0.013007111019558378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,2560,0.0069351109365622205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,512,0.0058622223635514574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,6144,0.009902222288979424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,7168,0.010378667049937779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,12288,0.02770311137040456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,6144,0.25369244151645237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,4096,0.007668444679843054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,3072,0.031806220610936485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,2048,0.006207110981146495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,128,0.005116444494989184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,2048,0.005399111244413588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,768,0.017176000608338248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,32,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,128,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,512,0.05480800072352091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,8192,0.016837333639462788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,5120,0.029122667180167303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,16384,0.019944889677895438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,2560,0.007678222325113084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,128,0.014752000570297241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,1024,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,3072,0.028816888729731243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,2560,0.013755555782053204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,7168,0.004588444613748127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,256,0.015504888362354703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,7168,0.009703999592198266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,256,0.0022524444179402455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,3072,0.007931555310885111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,3072,0.03767377800411648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,128,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,512,0.015831111205948722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,32,0.001830222292078866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,5120,0.02069244450993008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,768,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,64,0.002191999927163124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,256,0.015689777003394235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,1536,0.0661262207561069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,256,0.005879999862776862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,32,0.002830222249031067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,1024,0.07988444301817152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,128,0.008346666892369589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,4096,0.014505777094099255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,8192,0.010016000105275048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,10240,0.02012800011369917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,65536,0.17834044827355278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,5120,0.057641777727339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,32,0.006288888967699475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,768,0.0028577778074476454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,7168,0.009927999642160203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,768,0.02292177743381924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,5120,0.032476445039113365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,8192,0.264320002661811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,256,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,32,0.002529777793420686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,2560,0.012426666915416718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,65536,0.4379306634267171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,4096,0.012275555895434486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,65536,0.06152355670928955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,32,0.013413333230548434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,7168,0.009830222361617619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,1536,0.0063288890653186375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,1536,0.0035617777870761026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,2048,0.04029511080847846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,256,0.01407999959256914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,6144,0.009992000129487779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,4096,0.035029331843058266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,65536,2.7790968153211804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,5120,0.025755556093321905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,5120,0.10621333122253418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,2560,0.010738666686746808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,65536,0.17941866980658638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,1536,0.009654222263230218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,2560,0.01313066648112403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,1024,0.041781332757737905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,4096,0.004520888957712385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,3072,0.015284443895022074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,8192,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,128,0.002191999927163124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,3072,0.015823110938072205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,512,0.01316088851955202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,256,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,128,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,2048,0.014691554837756686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,768,0.016428444120619033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,128,0.01072533345884747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,32,0.008363555702898238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,1536,0.026132444540659588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,10240,0.0351839992735121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,2048,0.01054844425784217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,16384,1.3304017384847004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,12288,0.045240001546012036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,256,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,12288,0.018731556004948087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,256,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,12288,0.06310400035646227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,1024,0.004505777938498391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,3584,0.004884444591071871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,65536,0.7406257523430718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,6144,0.03882666760020786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,128,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,16384,0.015111999379263984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,4096,0.006552888701359431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,8192,0.014459555347760519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,1536,0.025479111406538222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,12288,0.03703555464744568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,1536,0.003984000119898054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,64,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,6144,0.013739555246300168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,1024,0.016187555260128446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,1536,0.011075555450386472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,512,0.005919111271699269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,2560,0.004922666483455234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,1024,0.025451555848121643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,768,0.0038986665507157645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,1536,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,768,0.017196445001496207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,2560,0.006636444479227066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,64,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,32,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,4096,0.011021333436171213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,8192,0.008151999778217739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,3584,0.016050666570663452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,65536,0.0829662217034234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,16384,0.02490577763981289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,256,0.007962666451931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,6144,0.06285778019163343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,1024,0.00812000036239624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,10240,0.009356444080670675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,8192,0.01274311128589842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,1024,0.01484444406297472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,2560,0.07739110787709554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,3584,0.007271111011505127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,1536,0.013773333695199756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,8192,0.010932444698280759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,3072,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,7168,0.006583999842405319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,7168,0.014135999812020196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,65536,0.23485244645012748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,10240,0.04285066657596164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,7168,0.023401778605249193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,128,0.007954667011896769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,5120,0.018961777289708454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,64,0.002163555473089218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,12288,0.013927110367351107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,64,0.007248889240953658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,6144,0.030903998348448012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,2048,0.007262222468852997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,512,0.005192000004980299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,65536,2.1564773983425565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,3072,0.011035555766688453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,3584,0.018567999203999836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,8192,0.024272888898849487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,8192,0.020940444535679285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,6144,0.02056533263789283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,16384,0.16311821672651503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,128,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,256,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,4096,0.007736889024575551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,2560,0.007356444166766272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,1536,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,16384,0.019911999503771465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,8192,0.03050133254792955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,8192,0.015308444698651632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,6144,0.01717422240310245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,2048,0.03156177865134345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,1536,0.006585777633719974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,512,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,12288,0.008311111066076491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,5120,0.008887999587588841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,3072,0.030914667579862807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,2560,0.033032000064849854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,8192,0.01514844430817498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,7168,0.17749244636959502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,8192,0.518655088212755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,256,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,65536,0.09529599878523086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,7168,0.04308533337381151
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,65536,0.054818666643566556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,128,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,10240,0.062103999985588916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,12288,0.019190222024917603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,128,0.00499288903342353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,2048,0.013452444639470843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,128,0.0069120000633928515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,512,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,1024,0.007666666474607255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,4096,0.011172444456153445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,32,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,768,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,8192,0.21179821756150988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,128,0.009901333186361525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,128,0.00591111100382275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,1024,0.004648000001907349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,8192,0.017668444249365065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,12288,0.026378666361172993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,3584,0.021057777934604224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,512,0.0038933331767717996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,16384,0.1659324434068468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,5120,0.009292444421185387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,768,0.006638222270541721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,12288,0.05120177732573616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,3584,0.013426666458447775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,2560,0.00562755556570159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,7168,0.0483511123392317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,8192,0.020799999435742695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,64,0.006568000134494569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,128,0.006274666637182236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,10240,0.010915555887752108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,32,0.00553955551650789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,65536,0.09000888797971938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,10240,0.03386666708522373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,3072,0.006959999601046245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,3584,0.018069333500332303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,256,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,512,0.08594222201241387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,64,0.011054221954610614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,8192,0.03040177623430888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,256,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,2048,0.004599110947714912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,2048,0.02162844439347585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,6144,0.03224000003602769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,7168,0.19198666678534615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,512,0.005214222189452913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,2560,0.016417778200573392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,4096,0.007361777954631382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,512,0.020079111059506733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,2048,0.013053333593739403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,10240,0.03515999846988254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,3584,0.023988443944189284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,768,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,10240,0.05211822191874186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,256,0.0030337776988744736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,10240,0.01144266708029641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,7168,0.017694221602545846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,1536,0.005971555494599872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,768,0.016522667474216886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,16384,0.07944977945751615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,4096,0.008284444610277811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,2048,0.005051555732885997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,3584,0.007689778175618913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,1024,0.007308444215191736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,256,0.017854221993022494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,5120,0.008669333325492011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,1024,0.006204444501135085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,128,0.01607022186120351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,128,0.011509333219793109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,3584,0.010031999813185798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,7168,0.009704888694816166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,128,0.008967999782827165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,32,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,12288,0.02198666665289137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,16384,0.023481777972645227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,256,0.007981333467695449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,7168,0.009724444813198512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,256,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,2048,0.006632888896597757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,256,0.0026853332916895547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,7168,128,0.018328888548745047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,2560,0.021340444684028625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,2048,0.008072888685597314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,8192,0.01038933296998342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,4096,0.00795288880666097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,32,0.005285333428117964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,1024,0.021993777818149988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,1536,0.005950222412745158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,16384,0.009674666656388177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,5120,0.011006222003036074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,5120,0.027059555053710938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,128,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,3072,0.014190221826235453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,2048,0.07332089212205675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,6144,0.008977777428097194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,5120,0.016175111134847004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,64,0.0021271111650599376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,65536,0.14576178126864964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,2048,0.009452444811662039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,4096,0.01777777738041348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,1024,0.01752355529202355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,8192,0.0188719994492001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,3584,0.011025778121418424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,768,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,16384,0.028452446063359577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,5120,0.0047208890318870544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,1024,0.013783111340469785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,1536,0.006993778049945831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,64,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,2048,0.02488888965712653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,16384,0.015847111741701763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,8192,0.03215822246339586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,3584,0.10203111171722412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,1024,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,256,0.005682666682534748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,8192,0.031558222240871854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,65536,0.18223999606238472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,512,0.027425777581002977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,16384,0.4042079978519016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,3584,0.00796977761718962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,128,0.0021262222694026097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,8192,0.05103288756476509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,8192,0.013456000222100152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,12288,0.018036444981892902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,1536,0.021340444684028625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,10240,0.02388622197839949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,12288,0.03723377651638455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,512,0.005138666679461797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,16384,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,512,0.020883555213610332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,12288,0.2700248824225532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,512,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,7168,0.13264977931976318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,32,0.002130666747689247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,6144,0.008352888955010308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,1024,0.012288000020715924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,10240,0.011346666349305047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,256,0.014429334137174817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,1536,0.11162400245666504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,128,0.014480888843536377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,768,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,6144,0.09248266617457072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,128,0.005199111170238919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,16384,0.00553688903649648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,10240,0.007532444265153673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,2560,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,7168,0.018536888890796237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,3072,0.015616888801256815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,65536,0.027093332674768236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,16384,0.3922124438815647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,16384,0.059584889147016734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,256,0.012620444099108378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,32,0.006646222124497096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,65536,0.06968533330493502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,10240,0.016128889388508268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,512,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,5120,0.008515555825498369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,2560,0.008079110748238033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,3584,0.005538666827811136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,6144,0.01648977730009291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,1536,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,128,0.002033777741922273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,2048,0.006736889067623351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,768,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,768,0.005791999813583162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,6144,0.056927111413743764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,4096,0.004655999855862724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,256,0.0022248888595236675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,3584,0.037547555234697126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,12288,0.02357599967055851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,4096,0.010779555473062726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,65536,1.2278969022962782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,2560,0.006718222051858902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,2048,0.01016355554262797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,1536,0.0038373333712418876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,6144,0.029078221983379785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,65536,0.06483644247055054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,3072,0.01254666679435306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,6144,0.027859555350409612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,2048,0.006238222122192383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,32,0.003863111138343811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,8192,0.03118844495879279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,1536,0.003559111307064692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,5120,0.013613333304723104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,5120,0.022664889693260193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,2560,0.039148443275027804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,16384,0.050957332054773964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,64,0.0018595556418100994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,8192,0.016774222254753113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,3584,0.011134222149848938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,128,0.01921599937809838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,128,0.005199111170238919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,1024,0.006229333165619109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,3584,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,768,0.005333333379692501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,256,0.005048888838953442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,768,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,5120,0.008623110751310984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,12288,0.01996888882584042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,1536,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,768,0.013240888714790344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,2048,0.026102221674389307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,2048,0.00628977765639623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,3072,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,768,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,128,0.012659555508030785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,3072,0.05208977725770739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,10240,0.7541644308302137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,12288,0.02021511064635383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,5120,0.008641777767075432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,6144,0.011694221860832639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,4096,0.021951110826598272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,2560,0.014426666829321118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,3072,0.14861778418223062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,10240,0.012442666623327466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,5120,0.02090666691462199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,256,0.008300444318188561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,1024,0.01721333298418257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,8192,0.02574311031235589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,128,0.005706666658322017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,64,0.005596444424655702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,5120,0.17781689431932238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,3584,0.007652444144090016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,3072,0.025245333711306255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,12288,0.054478221469455294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,3072,0.017855111095640395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,4096,0.016760889026853774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,5120,0.010024000373151567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,12288,0.018887110882335238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,65536,0.8894391589694552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,32,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,2048,0.6525120205349392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,64,0.005542222410440445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,1536,0.005967111223273807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,8192,0.014615999327765571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,768,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,768,0.003181333343187968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,1536,0.014860444598727755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,16384,0.0359866652223799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,16384,0.03407111101680332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,512,0.008325333396593729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,3584,0.021720889541837905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,5120,0.026132444540659588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,65536,0.8713679843478732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,8192,0.01109333336353302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,1536,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,10240,0.019699555304315355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,2048,0.012576888832781049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,128,0.002177777803606457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,512,0.020611556039916146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,768,0.0038986665507157645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,2560,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,2560,0.14757333861456975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,8192,0.09187466568417019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,8192,0.019308444526460435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,512,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,64,0.0063182223174307085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,4096,0.014856000741322836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,2560,0.019111111760139465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,2560,0.006591111007663939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,256,0.016150222884284127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,2048,0.018911111685964797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,3072,0.006976889239417181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,32,0.0018577778504954444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,512,0.014360888136757744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,16384,0.09948266877068414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,2560,0.004965333475006951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,6144,0.019877332780096266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,128,0.0032746667663256326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,16384,0.019448889626397025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,12288,0.2893199920654297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,5120,0.016894222961531747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,3072,0.029525332980685767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,2560,0.006984000404675801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,32,0.0052586665583981415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,5120,0.011853333148691388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,8192,0.011066666907734342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,3584,0.04792444573508369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,512,0.006016888966162999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,3584,0.06602133644951715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,2560,0.03323466579119364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,10240,0.015469332536061605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,12288,0.9480204052395291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,512,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,16384,0.32750844955444336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,768,0.006309333360857434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,1024,0.0031662223239739737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,5120,0.009317333499590555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,128,0.002157333410448498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,4096,0.0343537794219123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,2560,0.042437334855397545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,12288,0.02291822267903222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,16384,0.10169777605268691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,4096,0.01683999929163191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,12288,0.012736000120639801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,32,0.0373564428753323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,5120,0.01204177737236023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,2560,0.007215110792054071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,32,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,2560,0.010394666757848527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,512,0.013440888788965015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,4096,0.04779111014472114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,768,0.0031715554909573663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,2560,0.007146666447321574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,3584,0.01640888883007897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,12288,0.026743110683229234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,3584,0.022301332818137273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,512,0.004936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,1536,0.021947556071811255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,3072,0.007520889242490132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,5120,0.028781334559122723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,7168,0.07162755727767944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,10240,0.4895671208699544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,32,0.0021333334346612296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,1024,0.007385777930418651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,12288,0.02031822171476152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,2560,0.0069431112044387394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,7168,0.008755555583371056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,256,0.006629333313968446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,1024,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,2560,0.04717511269781324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,64,0.002986666642957263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,32,0.0025271111064487034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,3584,0.09152622355355157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,65536,0.020933333370420668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,5120,0.008638222184446123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,7168,0.013422222601042854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,10240,0.0727528863483005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,2048,0.0079644446571668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,256,0.01069866700304879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,3072,0.009518221848540837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,10240,0.08866222037209405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,2048,0.006618666566080517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,12288,0.009328000247478485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,256,0.004862222406599256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,128,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,128,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,65536,0.06534755229949951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,3072,0.007626666790909237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,32,0.002817777709828483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,1024,0.011748444702890186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,4096,0.013425778183672162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,2560,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,512,0.00794400026400884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,2560,0.007311999797821045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,6144,0.021985777550273474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,3584,0.005106666849719154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,32,5120,0.004860444615284602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,64,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,1024,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,4096,0.015621332658661736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,256,0.00250133333934678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,2560,0.021728888154029846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,1024,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,2560,0.012832889126406776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,8192,0.024826667375034753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,10240,0.009512888888518015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,65536,0.082096000512441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,6144,0.01442399952146742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,8192,1.204742219712999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,5120,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,7168,0.018751111295488145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,3584,0.017858665850427415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,10240,0.16538488864898682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,7168,0.009704888694816166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,3584,0.06862666871812609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,6144,0.038973334762785174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,3072,0.006940444310506185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,512,0.004375111311674118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,2048,0.02089511189195845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,4096,0.012450666891203986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,1536,0.006313777632183499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,6144,0.03171022070778741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,64,0.002507555608948072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,16384,0.29340889718797475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,64,0.008051555189821456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,1024,0.022308443983395893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,16384,0.035544888840781316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,1024,0.006266666783226862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,3072,0.14610221650865343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,6144,0.022096888886557683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,512,0.01940533353222741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,768,0.013068444199032254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,2048,0.006301333506902059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,8192,0.025311110748185053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,32,0.002143111079931259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,1536,0.005976000179847081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,3584,0.01717155509524875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,65536,0.4253999922010634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,7168,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,8192,0.015054222610261707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,10240,0.021708443760871887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,1536,0.018223999275101554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,10240,0.3355315526326497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,3072,0.007589333587222629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,256,0.007832888927724626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,7168,0.020037333170572918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,2048,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,3072,0.022696000006463792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,1024,0.017887999614079792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,8192,0.021320889393488567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,8192,0.09118577506807114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,768,0.007273777491516537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,16384,0.021470222208235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,7168,0.01181422256761127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,1024,0.017979555659823947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,12288,0.028543111350801256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,10240,0.019647111495335896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,512,0.004600888739029567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,3584,0.020658666888872784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,512,0.004942222187916438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,12288,0.065792891714308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,7168,0.015122666954994202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,2560,0.014469333820872836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,3584,0.02019466625319587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,8192,0.005615111026499007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,7168,0.029341333442264136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,7168,0.01713244451416863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,64,0.005201777650250329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,1536,0.014428445034556918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,128,0.051685333251953125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,6144,0.034280889564090304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,3072,0.015684444043371413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,4096,0.02383822202682495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,128,0.004879999905824661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,128,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,64,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,512,0.016169778174824186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,512,0.0024951110697454875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,768,0.007663999994595845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,7168,0.016707556115256417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,6144,0.008972444468074376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,3584,0.00794577764140235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,12288,0.017867555220921833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,5120,0.028372443384594385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,16384,0.016891555653678048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,512,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,2048,0.006308444258239534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,6144,0.018230222993426852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,65536,0.10582844416300456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,12288,0.04858044452137417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,1536,0.016900445024172466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,3072,0.010346666806273991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,65536,0.269399987326728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,3584,0.019890666007995605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,1536,0.006071110980378256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,12288,0.039106667041778564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,3072,0.008654221892356873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,7168,0.010666666759385003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,16384,0.06977422369851007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,65536,0.11698933442433675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,3072,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,2560,0.03380889031622145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,1536,0.007187555233637492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,3072,0.011029333704047732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,7168,0.011751111182901593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,2048,0.007709333466158972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,512,0.015776889191733468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,32,0.0018026667336622875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,12288,0.09007377756966485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,1536,0.018498667412334018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,12288,0.011387555963463254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,512,0.0031822222388452957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,256,0.0028524444335036804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,32,0.09044444561004639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,128,0.014390221900410123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,1024,0.008797333472304875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,4096,0.007342221836249034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,2048,0.008631111019187504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,4096,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,7168,0.015826667348543804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,1024,0.006237333433495627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,2560,0.006250666661394968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,512,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,256,0.007727111379305522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,16384,0.024541333317756653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,8192,0.019355555375417072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,1024,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,1536,0.006073777874310811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,3584,0.005965333431959152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,16384,0.03547022077772353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,32,0.00282577777074443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,1024,0.004934222333961063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,256,0.004920000003443824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,768,0.0038417776425679526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,5120,0.019162666466501024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,256,0.004874666531880696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,64,0.005633777628342311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,7168,0.12725422117445204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,3072,0.02778488728735182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,1024,0.027820444769329492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,1024,0.005646222167544895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,256,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,6144,0.03295911020702786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,8192,0.016844444804721408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,2048,0.006862222320503659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,16384,0.10911199781629775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,7168,0.026771555344263714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,5120,0.014117333624098035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,512,0.0053377776510185665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,768,0.017867555220921833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,1536,0.009343999955389235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,512,0.011983999775515662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,2048,0.049399998452928334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,8192,0.09171111053890652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,12288,0.024051555328898962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,64,0.00797333319981893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,4096,0.01790844400723775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,128,0.014465777410401238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,768,0.0029306666304667792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,128,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,3072,0.9453421698676215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,5120,0.011711111499203576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,2048,0.0309804446167416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,128,0.040404445595211454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,3072,0.016270221935378183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,128,0.011605333122942181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,12288,0.03439022103945414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,128,0.0019786666250891155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,256,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,16384,0.0725404421488444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,768,0.005544888890451855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,16384,0.034055110481050276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,256,0.03589866558710734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,256,0.015802666544914246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,65536,0.1392337746090359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,6144,0.11901777320437962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,6144,0.03851999839146932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,2560,0.05155377917819553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,4096,0.03086400032043457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,5120,0.015454222758611044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,7168,0.01753777762254079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,2560,0.012255999777052136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,6144,0.03756533397568597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,256,0.00499288903342353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,8192,0.014173333843549093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,256,0.014532445205582513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,5120,0.008948444492287105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,256,0.00249777775671747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,768,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,3584,0.04267377654711405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,10240,0.01461777753300137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,3584,0.004546666724814309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,5120,0.008739555875460306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,5120,0.0123831108212471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,2560,0.006601777755551868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,65536,0.09596800141864353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,3072,0.0236817780468199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,768,0.04810933272043864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,256,0.013401777380042605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,2048,0.0038497779104444715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,1024,0.0052168890833854675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,3584,0.009007111191749573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,16384,0.027922666735119287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,3584,0.03282489048110114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,256,0.029475556479560003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,2560,0.007055111229419708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,2560,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,1024,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,128,0.0028168888141711554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,256,0.013702222042613559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,4096,0.059168888462914355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,7168,0.019208888212839764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,1536,0.018192888961897958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,8192,0.010687111152542962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,2560,0.006923555499977536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,10240,0.011374221907721626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,256,0.005530666559934616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,3584,0.007230222225189209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,512,0.018062222335073683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,2048,0.00656088896923595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,10240,0.17480088604821098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,10240,0.020845333735148113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,5120,0.01552088889810774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,4096,0.10763911406199138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,2048,0.004581333448489507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,16384,0.027657777070999146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,3584,0.027249778310457867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,7168,0.019162666466501024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,768,0.0069413334131240845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,4096,0.016209777858522203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,768,0.020640888147883944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,10240,0.02548800077703264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,7168,0.03353155652681986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,10240,0.06910133361816406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,5120,0.013224889006879596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,512,0.002508444504605399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,6144,0.012479999827014076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,512,0.003177777760558658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,1536,0.0075528886583116316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,64,0.00519822206762102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,8192,0.04057155715094672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,10240,0.04958577619658577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,1536,0.006241777704821692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,16384,0.005720888988839255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,2560,0.015189333094490899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,4096,0.032953778902689614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,6144,0.028519110547171697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,1536,0.003592888928121991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,2048,0.008336889247099558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,16384,0.03736977775891622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,2048,0.004577777865860197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,1024,0.004515555583768421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,16384,0.02571022179391649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,3072,0.014170666535695394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,8192,0.01514755520555708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,3072,0.0148053334818946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,4096,0.3468080096774631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,256,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,16384,0.015141333142916361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,768,0.017159110969967313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,64,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,12288,0.022775999373859827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,1536,0.011861333416567909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,256,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,1536,0.013295999831623502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,65536,0.18430844942728677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,256,0.00563733321097162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,16384,0.04493511054250929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,4096,0.006574222197135289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,3584,0.010777778095669217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,2560,0.36446311738755965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,12288,0.019270223047998216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,768,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,10240,0.023709333605236475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,128,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,7168,0.010074666804737514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,3072,0.027443556321991816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,12288,0.07312444183561537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,768,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,2560,0.014857777290874057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,65536,4096,1.2230293485853407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,128,0.017204445269372728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,2048,0.006996444529957241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,1536,0.013414222333166333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,512,0.016568889220555622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,16384,0.03163999981350369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,10240,0.037363555696275495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,65536,1.28583738538954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,128,0.008665777742862701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,6144,0.08524177471796672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,128,0.0025155554629034465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,256,0.015269334117571512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,12288,0.051896890004475914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,16384,0.018840889135996502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,128,0.009005332986513773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,512,0.0029804443733559716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,2560,0.065102219581604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,2560,0.012703110774358114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,1536,0.005976888868543837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,3584,0.008847110801272923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,32,0.0021804444905784395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,3584,0.06994933552212186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,3584,0.012192000117566852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,4096,0.01089511149459415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,12288,0.02745955520206028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,256,0.010020444790522257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,1536,0.005923555543025334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,256,0.004884444591071871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,7168,0.012056888805495368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,3584,0.05496889021661547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,3072,0.0933297806315952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,128,0.006629333313968446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,6144,0.011395555403497485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,7168,512,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,1536,0.05504711137877571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,6144,0.020056888461112976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,10240,0.023557333482636347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,2048,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,10240,0.16702577802870008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,2048,0.060452441374460854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,5120,0.07678666379716662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,64,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,7168,0.004884444591071871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,6144,0.05829422341452705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,65536,0.20773777696821424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,2560,0.01202044470442666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,3584,0.007611555357774098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,2560,0.01092977739042706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,128,0.004880889008442561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,65536,0.1982426643371582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,2048,0.008661333057615492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,2560,0.0069644442862934535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,6144,0.012231999801264869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,16384,0.20315910710228813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,512,0.006516444600290722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,3072,0.006778666542636023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,1024,0.07290399736828275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,32,0.006268444574541516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,8192,0.01036977767944336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,5120,0.013977777626779346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,4096,0.011570666399266986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,128,0.014133334159851074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,128,0.014135999812020196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,7168,0.049504889382256397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,1024,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,65536,0.07679644558164808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,128,0.01891644464598762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,16384,0.018815999229749043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,128,0.01108266661564509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,256,0.016177778442700703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,2560,0.02160088883505927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,3584,0.007299555672539606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,32,0.00590844452381134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,4096,0.027373333772023518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,5120,0.011784888803958893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,1536,0.007690666450394525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,768,0.0025111111915773815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,6144,0.10968977875179714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,32,0.013383111192120446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,7168,0.007284444239404466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,1536,0.025796444879637823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,16384,0.02513244416978624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,12288,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,512,0.05499644411934746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,3072,0.009362666971153682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,1024,0.023377777801619634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,4096,0.007948444121413762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,65536,0.13852532704671225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,2048,0.054735110865698926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,256,0.014045332868893942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,10240,0.012434666355450949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,3072,0.03775555557674832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,256,0.01775644388463762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,5120,0.03801866703563266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,1024,0.01583377851380242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,16384,0.01794666714138455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,5120,0.24776355425516763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,5120,0.02678666677739885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,5120,0.014764444695578681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,2560,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,2048,0.006279110908508301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,128,7168,0.020351111888885498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,65536,0.0413351125187344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,1024,0.007354666789372762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,256,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,7168,0.01886133352915446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,1024,0.005604444278611078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,10240,0.023429332507981196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,256,0.005886222339338726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,10240,0.023031999667485554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,128,0.002463999928699599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,2048,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,7168,0.17034044530656603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,16384,0.02602044410175747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,2048,0.011776000261306763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,6144,0.4878080156114366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,4096,0.011977777712874942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,768,0.00699733363257514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,7168,0.020954666866196528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,7168,0.027738667196697656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,12288,0.019707555572191875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,12288,0.02917333443959554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,128,0.012392889294359418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,32,0.003216000066863166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,2048,0.018761777215533786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,1536,0.0036959999965296853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,512,0.002575111058023241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,128,0.010354667074150508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,3584,0.004851555658711327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,8192,0.011050666371981302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,768,0.04208799865510729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,10240,0.03548444310824076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,3072,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,16384,0.05780888928307427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,6144,0.011360888679822287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,5120,0.0041964443193541635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,65536,0.07728888591130574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,6144,0.005663110978073544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,2048,0.007111111448870764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,2048,0.00942844483587477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,2048,0.006242666807439592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,2560,0.0069004446268081665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,16384,0.024118221468395654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,64,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,8192,0.004881777697139316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,768,0.015460444821251763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,10240,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,8192,0.014446222119861178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,1536,0.0064897777305708984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,64,0.0018311111877361934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,12288,0.027148443791601393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,12288,0.024960888756646052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,5120,0.020738666256268818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,12288,0.034847110509872437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,4096,0.007729777859316931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,2560,0.18917866547902426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,3584,0.015110222829712762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,6144,0.014500444134076437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,8192,0.013109332985348172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,256,0.004933333231343163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,6144,0.20888888835906982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,5120,0.019529778096410964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,1024,0.015142222245534262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,1536,0.006329777754015393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,4096,0.025635555386543274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,2560,0.01484444406297472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,1536,0.008192000289758047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,8192,0.04715111189418369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,4096,0.005655111124118169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,3072,0.005583111196756363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,32,0.005888000130653381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,2560,0.00846400029129452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,5120,0.011687111523416309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,10240,0.020240000552601285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,10240,0.05399644374847412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,32,0.002528000002106031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,128,0.007690666450394525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,2048,0.013904889424641928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,5120,0.026155556241671245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,512,0.010088889135254754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,16384,0.009653333160612319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,12288,0.04607111215591431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,1536,0.0035084444615576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,5120,0.01458044515715705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,16384,0.023668444818920557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,16384,0.05709866682688395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,256,0.006653333289755716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,3072,0.01756622228357527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,512,0.01549600064754486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,6144,0.05751644240485298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,7168,0.009639110830095079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,1024,0.00792266676823298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,65536,0.07833421892589994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,2560,0.010455110834704505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,5120,0.012703110774358114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,1024,0.014862222803963555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,6144,0.03125066558519999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,256,0.02717511190308465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,4096,0.012072888513406118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,768,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,12288,0.025006221400366888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,7168,0.012484444512261285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,6144,0.01789066692193349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,7168,0.02271911170747545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,256,0.01443555619981554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,3072,0.03529955612288581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,8192,0.0551457769340939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,32,0.010041777458455827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,512,0.009683555199040307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,2048,0.01310400002532535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,12288,0.02017866737312741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,3072,0.009776888622177972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,2560,0.011981333295504252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,2048,0.006217777729034424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,16384,0.007650666766696506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,8192,0.1515671147240533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,256,0.0058053334554036455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,1536,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,7168,0.028990222348107234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,6144,0.03870044483078851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,256,0.01331822243001726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,256,0.014462222655614218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,512,0.010866666833559671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,8192,0.18674311372968885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,1024,0.009081777599122789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,128,0.004961777892377642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,512,0.002495999965402815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,65536,0.10678666830062866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,2560,0.01239911135700014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,10240,0.3450755543178982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,12288,0.05079288946257698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,1536,0.022617777188618977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,3072,0.015486222174432544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,1024,0.03567466802067227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,2048,0.023039110832744177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,512,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,3584,0.0079644446571668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,512,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,12288,0.01408622165520986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,1024,0.005607999861240387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,12288,0.012704888979593912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,65536,0.1467991140153673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,8192,0.01368799971209632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,3584,0.2683546543121338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,5120,0.009000000026490953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,768,0.022887999812761944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,16384,0.023781332704755995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,2048,0.04745955599678887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,1024,0.00423200014564726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,3584,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,3584,0.02762311034732395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,10240,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,128,0.021663111117151048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,65536,0.5462613105773926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,64,0.002032888846264945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,128,0.002185777864522404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,512,0.016499555773205228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,3072,0.0041857775714662345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,1536,0.1195946666929457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,1536,0.005919111271699269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,7168,0.03498933381504483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,768,0.003168888803985384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,4096,0.006625777731339137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,16384,0.0864222248395284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,32,0.006265777680608962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,2048,0.05290311243798998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,3584,0.007599111232492659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,1536,0.0038044444388813446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,12288,0.013054221868515015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,7168,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,8192,0.023898666103680927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,2560,0.03219022353490194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,5120,0.0151582227812873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,1024,0.005967111223273807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,64,0.001830222292078866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,8192,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,768,0.003141333245568805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,32,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,768,0.002887111157178879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,8192,0.05832799937989977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,3584,0.064173334174686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,3072,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,12288,0.009126221968068017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,768,0.01754577789041731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,6144,0.013758222262064615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,6144,0.017477333545684814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,2560,0.021645334031846788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,3072,0.012718222207493253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,1536,0.005961777849329843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,256,0.00217333332531982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,16384,0.026872888207435608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,12288,0.01639111174477471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,16384,0.708373334672716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,1536,0.007298666569921706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,65536,0.06600177950329252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,3072,0.016975111431545682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,3072,0.00924622184700436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,2560,0.011776000261306763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,2048,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,4096,0.011721777419249216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,4096,0.016538666354285348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,2560,0.055065777566697865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,65536,0.1517582204606798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,1024,0.014231999715169271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,10240,0.03080000148879157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,1024,0.002879111096262932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,4096,0.00684799998998642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,1024,0.005533333453867171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,6144,0.008629333641793992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,1536,0.007592000067234039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,512,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,16384,0.014783999986118741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,1536,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,1024,0.011378666592968835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,1536,0.006261333409282897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,5120,0.008599110775523717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,3072,0.007676444119877285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,512,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,3072,0.014833778142929077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,16384,0.017272000511487324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,3584,0.011612444288200803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,3584,0.13387289312150744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,2048,0.007722666694058313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,4096,0.055693335003323026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,512,0.05270488725768196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,3584,0.023669333921538457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1024,3072,0.014488889111412896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,8192,0.0229759994480345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,8192,0.02330844435426924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,128,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,2048,0.0070879997478591064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,2560,0.007090667055712805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,2560,0.01825066738658481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,768,0.032668444845411516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,65536,0.060439997249179415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,12288,0.018215111560291715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,5120,0.016930666234758165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,16384,0.055867552757263184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,32,0.009317333499590555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,5120,0.028723554478751287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,6144,0.020573332905769348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,6144,0.017151999804708693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,3584,0.010325333310498131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,64,0.0018088887963030075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,512,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,5120,0.02490933405028449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,4096,0.022291556000709534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,2560,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,768,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,1536,0.006230222268237009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,256,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,3072,0.014114666316244336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,3072,0.006959999601046245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,2048,0.007266666326257918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,32,0.009650666680600908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,16384,0.019546666079097323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,4096,0.008283555507659912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,10240,0.03800800111558702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,12288,0.01927288870016734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,512,0.05438399977154202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,256,0.0027715555495686005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,7168,0.013739555246300168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,2048,0.06692533360587226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,8192,0.01572711103492313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,7168,0.025672889418072168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,2048,0.021151110529899597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,7168,0.009661333428488838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,16384,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,10240,5120,0.28504355748494464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,1024,0.005892444401979446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,6144,0.008965333302815756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,2560,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,16384,0.02827822168668111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,12288,0.005579555614127054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,5120,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,3584,0.02458311120669047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,768,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,512,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,1024,0.007271111011505127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,5120,0.00850222259759903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,7168,0.03019733230272929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,5120,0.03684088918897841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,10240,8192,0.04029688901371426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,32,0.005320888840489917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,10240,0.01426488823360867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,3072,0.004589333302444882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,256,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,12288,0.01716088917520311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,5120,1.0375341839260526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,512,0.015027556154463025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,2560,0.008458666503429413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,1024,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,16384,0.004556444370084339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,512,0.005623111294375525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,4096,0.01995022263791826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,2048,0.07102844450208876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,2560,0.029337777031792536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,64,0.002458666761716207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,10240,0.03760622276200189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,12288,0.05857866340213352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,8192,0.17431378364562988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,7168,0.012054222325483957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,3584,0.016545777519543965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,3072,0.014632888966136508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,8192,0.016226665841208566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,7168,0.020564445190959506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,2048,0.006215111249023014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,2048,0.0041786668201287585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,2560,0.02738933265209198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,10240,0.20315910710228813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,512,0.0041706665522522396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,7168,0.02218222286966112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,8192,0.031158222092522517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,16384,0.02572888963752323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,4096,0.03434755404790243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,32,0.006607999818192587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,4096,0.011386666860845355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,6144,0.010040000081062317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,1536,0.015133332875039844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,8192,0.032356444332334734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,2048,0.024597333537207708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,12288,0.013510222236315409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,12288,0.016598222984208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,8192,0.011086222198274402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,2048,0.0063279999627007385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,12288,0.30841154522365993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,1536,0.0063164445261160536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,6144,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,3584,0.016142222616407607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,128,0.0069439998931354946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,768,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,256,0.0053013331360287136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,5120,0.015322667029168872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,8192,0.20762399832407633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,5120,0.009381333159075843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,5120,0.005382222019963794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,2048,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,12288,0.04890133274926079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,16384,0.0698382192187839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,16384,0.028404444456100464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,10240,0.0059075554211934405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,6144,0.026680888401137456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,7168,0.0192275560564465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,2048,0.006688000013430913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,12288,0.2192417780558268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,10240,0.06324533621470134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,1536,0.007981333467695449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,5120,0.016861332787407767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,256,0.002512000087234709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,5120,0.010784889260927836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,128,0.006573333508438534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,12288,0.01475644442770216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,3584,0.01036800030204985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,7168,0.005201777650250329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,12288,0.018552000323931377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,12288,0.1811279985639784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,16384,0.03829333186149597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,10240,0.029819554752773706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,512,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,768,0.015536889433860779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,2048,0.004514666481150521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,8192,0.025290666355027094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,8192,0.010671999719407825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,12288,0.03127911024623447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,8192,0.03743911120626662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,4096,0.008015111088752747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,2048,0.0069208890199661255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,8192,0.1486062208811442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,1536,0.004218666503826777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,128,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,64,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,4096,0.007875555919276344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,65536,0.18032622337341309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,512,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,1024,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,512,0.018254222141371835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,8192,0.03942844602796767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,2560,0.020249777370029025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,1024,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,16384,0.022388445006476507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,5120,0.018869333797030978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,7168,0.04511377877659268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,3072,0.015069334043396844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,4096,0.04519822200139364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,2048,0.02065955599149068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,1536,0.006105777704053455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,1024,0.008291555775536431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,768,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,128,0.0022053333620230355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,1024,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,64,0.0024684444069862366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,7168,0.031141334109836157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,65536,0.08835111061731975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,64,0.0020675555699401433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,65536,0.13600888517167833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,1536,0.018264888061417472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,10240,0.035232888327704534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,1536,0.016865778300497267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,1024,0.009335111412737105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,256,0.014726221561431885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,10240,0.017629333668284945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,2048,0.025378665990299646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,5120,0.015483554866578845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,2048,0.02439822256565094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,768,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,10240,0.02289155622323354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,2048,0.018544889158672757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,8192,0.0848204427295261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,1536,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,512,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,1536,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,16384,0.03028800090154012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,7168,0.0229759994480345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,10240,0.009349333743254343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,3072,0.015764445066452026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,1024,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,3584,0.007620444728268518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,1536,0.005863111052248213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,65536,0.236429320441352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,3072,0.02091111077202691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,1024,0.007640000018808577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,256,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,8192,0.02597155504756504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,1536,0.006351111249791251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,256,0.005209777918126848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,10240,0.0398595564895206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,2560,0.041377776198916964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,6144,0.02588711182276408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,512,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,10240,0.007537778053018782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,1536,0.005944889038801193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,1024,0.3231244352128771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,1024,0.005917333480384614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,256,0.0025048889219760895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,65536,0.18044977717929414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,128,0.01754666699303521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,512,0.18941156069437662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,768,0.00794044468137953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,8192,0.1648142205344306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,2560,0.007797333101431529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,8192,0.012724444270133972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,65536,0.08788800239562988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,3072,0.008558221989207799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,8192,0.10179199775060017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,8192,0.04359199934535556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,2560,0.008672888908121321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,6144,0.1427475611368815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,10240,0.012140444583363004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,768,0.014750222365061441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,2560,0.019549333386951022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,16384,0.03278755479388767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,3072,0.019283556275897555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,65536,0.1738826698727078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,7168,0.03293777836693658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,1536,0.014334221680959066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,32,0.017703110973040264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,2048,0.014848889576064216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,7168,0.014463999205165438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,8192,0.01962488889694214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,16384,0.04783199893103706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,128,0.013781333135233985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,12288,0.025639111797014873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,2560,0.03518666823705038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,2560,0.0052151112920708126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,1024,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,7168,0.15481066703796387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,3072,0.013380444712109037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,1024,0.02382933265633053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,256,0.0024826667375034755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,10240,0.04692177640067207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,128,0.004945777770545748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,128,0.009713778065310584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,128,0.0021253333737452826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,4096,0.011379555695586734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,8192,0.04022666811943054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,12288,0.09752977556652492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,768,4096,0.017532444662517972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,1024,0.00756888919406467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,6144,0.03982222080230713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,2048,0.004252444538805219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,65536,0.08610400226381089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,4096,0.011550222006109027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,64,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,256,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,10240,0.05007644494374593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,7168,0.010694222317801582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,128,0.01447111037042406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,7168,0.11975467205047607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,7168,0.013391999734772576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,1024,0.0068488890926043195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,3584,0.014831110835075378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,65536,0.015544889701737298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,16384,0.016626665989557903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,128,0.014118222726715935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,1024,0.11625066068437363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,16384,0.13844799995422363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,3584,0.02565866708755493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,5120,0.008630221916569604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,65536,0.0844915575451321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,512,0.010686222049925061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,8192,0.029037333197063867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,512,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,8192,0.02923911147647434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,16384,0.041998220814598926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,8192,0.010597333312034607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,3584,0.007302222152551015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,65536,0.47019375695122617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,1024,0.006908444480763541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,12288,0.024080889092551336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,10240,0.033370667033725314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,3072,0.007253333098358578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,16384,0.22901333702935112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,16384,0.016160888804329764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,1024,0.0031511110977994073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,64,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,7168,0.3507297833760579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,256,0.0029004443850782183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,1024,0.0058853332367208265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,7168,0.2073662281036377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,128,0.014451555079884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,6144,0.02598222262329525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,512,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,32,0.0026844443960322272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,512,0.01146755533085929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,5120,0.04255911045604282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,1024,0.01512888901763492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,1024,0.009312888814343346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,4096,0.041831112570232816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,12288,0.014329777823554145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,2560,0.01555022266176012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,7168,0.20844444963667128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,64,0.0021199999998013177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,7168,0.024098666177855596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,5120,0.011770666473441653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,65536,0.9053875605265299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,2048,0.033518221643235944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,2560,0.007083555890454187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,2048,0.007608888877762689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,10240,0.09307999743355645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,512,0.013773333695199756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,5120,0.03269511130121019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,128,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,12288,0.028762665059831407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,256,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,65536,0.24590577019585502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,7168,0.0292595558696323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,65536,0.07155644231372409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,10240,0.012442666623327466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,12288,0.019494222270117868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,128,0.0021928888228204516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,7168,0.01074399964676963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,256,0.005192000004980299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,1536,0.00388355553150177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,64,0.0064737776087390045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,512,0.01275644451379776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,128,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,8192,0.10336711009343465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,10240,0.2537902196248372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,65536,1.2264266543918185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,1536,0.009681777821646797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,512,0.014459555347760519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,4096,0.01518133282661438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,512,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,2560,0.01408622165520986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,65536,0.05436533358361986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,1536,0.027855111493004694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,3584,0.022289777795473736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,2048,0.013408000270525614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,32,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,3584,0.01613066593805949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,128,0.019314666589101154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,3584,0.007644444704055786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,32,0.035903112755881414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,8192,0.022290666898091633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,2048,0.016137777103318107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,3584,0.008956444760163626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,6144,0.019290667441156175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,128,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,64,0.002120888895458645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,2560,0.008173333273993598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,10240,0.023592000206311543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,7168,0.00778399987353219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,10240,0.023826667004161414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,3072,0.023060444328520033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,2048,0.022853333089086745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,4096,0.01681866745154063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,7168,0.010429333481523724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,3584,0.007598222129874759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,2560,0.006950221955776215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,4096,0.011060444845093621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,1024,0.0031537777847713898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,768,0.003192000091075897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,32,0.008113777471913232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,12288,0.03280177712440491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,16384,0.019896000623703003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,10240,0.016935999194780987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,7168,0.033779554896884494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,12288,0.019704000817404855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,2560,0.004914666629499859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,64,0.010984000232484607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,12288,0.016904888881577384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,128,0.01477688882086012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,8192,0.010651555326249866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,1024,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,64,0.0019902222686343724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,2560,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,3584,0.02752444479200575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,12288,0.04474755459361606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,1536,0.027802666028340656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,5120,0.0414684447977278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,7168,0.041033777925703264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,1536,0.009879999690585667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,1536,0.01889333294497596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,3072,0.006907555378145642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,1536,0.08620088630252415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,1536,0.016139555308553908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,10240,0.053731554084353976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,65536,0.3070862293243408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,2560,0.0069013333155049225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,7168,0.029550221231248643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,768,0.002883555574549569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,1536,0.01590133375591702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,4096,0.00425866660144594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,16384,0.020030222005314298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,64,0.008983111215962304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,12288,0.05125066637992859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,5120,0.052849776215023465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,8192,0.010687111152542962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,3584,0.008441777692900764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,256,0.01464355488618215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,16384,0.07398311297098796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,65536,0.07491999864578247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,1536,0.00351200004418691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,16384,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,7168,0.012723555167516073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,5120,0.00868444475862715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,1536,0.004471111214823193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,512,0.0021804444905784395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,768,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,12288,0.017256889078352187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,256,0.01036177741156684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,1024,0.0028204443968004654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,5120,0.018233777748213876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,2048,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,3072,0.01648977730009291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,1536,0.0038906666967603895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,10240,0.04488977789878845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,32,0.006640888750553131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,2048,0.007896888587209914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,2048,0.006296000132958095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,1024,0.003528000166018804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,3072,0.02444533341460758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,768,0.003119999956753519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,512,0.0052693333062860705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,64,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,1536,0.00796977761718962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,768,0.003192000091075897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,768,0.0038968887594011095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,12288,0.05137155453364054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,2048,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,10240,0.02158044444190131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,64,0.0031804444475306403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,3072,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,1536,0.006275555739800136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,128,0.004927111168702443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,16384,0.019204444355434842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,512,0.014494222071435718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,256,0.032933334509531655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,1024,0.005665777872006099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,1536,0.014879999889267815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,16384,0.04420800010363261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,4096,0.019107555349667866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,3584,0.04629866613282097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,3072,0.008735111190213097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,128,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,128,0.005205333232879639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,256,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,2560,0.009572444690598382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,65536,0.3606862227121989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,4096,0.029872000217437744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,256,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,128,0.006233777850866318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,64,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,3072,0.01036533299419615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,16384,0.020057777563730877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,1024,0.01551999979548984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,256,0.004559111264016894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,7168,0.00794933322403166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,8192,0.04434400134616428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,7168,0.01754577789041731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,2048,0.011952000359694162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,512,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,10240,0.03964711229006449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,3072,0.007986666427718269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,2560,0.010436444646782346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,10240,0.012127999630239276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,32,0.002130666747689247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,12288,0.01878666712178124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,65536,0.07501155800289579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,6144,0.010464000205198923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,12288,0.020958221620983545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,1024,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,1024,0.01683911184469859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,7168,0.015383111106024848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,5120,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,5120,0.01978933314482371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,3072,0.010010666317409938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,128,0.0018613334331247543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,12288,0.05998311440149943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,65536,0.2850408818986681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,512,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,1536,0.03824888997607761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,12288,2.1237199571397567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,64,0.00998044427898195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,10240,0.015134221977657743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,3584,0.019944889677895438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,10240,0.03138755427466498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,512,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,10240,0.05593422386381361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,7168,0.018883556127548218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,6144,0.08838399913575913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,12288,0.03552000059021844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,16384,0.11874932712978786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,65536,0.19535556104448107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,2048,0.006286222073766921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,2048,0.08760266833835179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,768,0.06150311231613159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,7168,0.023601777023739282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,64,0.005912000106440649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,4096,0.008691555923885768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,3072,0.007254222200976477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,10240,0.017870222528775532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,64,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,8192,0.016573333077960543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,7168,0.1668728854921129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,1024,0.0029084444459941653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,256,0.016186666157510545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,6144,0.026349332597520616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,2560,0.027451554934183758
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,16384,3.175361845228407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,65536,0.08305511209699842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,128,0.014496000276671516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,128,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,10240,0.0162808895111084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,65536,0.16385777791341147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,256,0.007656888829337225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,3584,0.011352000137170156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,10240,0.04191022117932638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,16384,0.022198221749729578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,7168,0.0138586668504609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,1536,0.009724444813198512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,4096,0.2750133408440484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,256,0.00665244460105896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,7168,0.014879110786649915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,16384,0.7217537562052408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,1024,0.010040000081062317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,2560,0.01997777819633484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,16384,0.03209688928392198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,3584,0.20315555731455484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,16384,0.05740977658165825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,128,0.007954667011896769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,8192,0.07332799832026164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,16384,0.03124177787039015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,2048,0.006262222097979651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,6144,0.013767110804716745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,512,0.007991111112965478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,256,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,7168,0.00963288876745436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,512,0.0075697774688402815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,12288,0.055664890342288546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,8192,0.02973244587580363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,128,0.014131555954615274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,1024,0.01482311056719886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,768,0.0041759999261962045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,512,0.002620444529586368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,2560,0.004926222066084544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,16384,1.1973110834757488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,3072,0.054867555697758995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,10240,0.01669066647688548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,1536,0.00657155571712388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,6144,0.01070666644308302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,65536,0.14036266009012857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,1024,0.005939555664857228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,1536,0.00795733349190818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,128,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,1024,0.01848266687658098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,10240,0.0247999992635515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,768,0.016528000434239704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,768,0.016178665889634024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,3072,0.018990221950742934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,2048,0.006734222173690796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,768,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,5120,0.006299555715587404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,768,0.016549333930015564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,32,0.00793511089351442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,1024,0.01484266585773892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,128,0.004888888862397936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,256,0.014848000473446317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,6144,0.011584888729784222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,512,0.02416977783044179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,3584,0.007304888632562425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,1536,0.02735111117362976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,1536,0.006258666515350342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,256,0.015475556254386902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,8192,0.054026666614744395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,3584,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,5120,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,5120,0.017842666970358956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,10240,0.004567999806669024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,3584,0.00792622235086229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,512,0.004907555464241239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,5120,0.008624888956546783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,2560,0.006486222147941589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,4096,0.0105262224872907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,128,0.01445777714252472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,256,0.006732444382376141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,6144,0.015060444672902426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,512,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,1024,0.007789333661397298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,3072,0.008351999852392409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,10240,0.07083821958965726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,16384,0.08631200260586208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,3584,0.011151110960377587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,256,0.016547555724779766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,5120,0.03730577892727322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,2048,0.006609777609507243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,1536,0.008616888688670265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,3584,0.037871999873055354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,4096,0.055270221498277455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,1536,0.02883466747072008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,65536,0.07482844591140747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,16384,0.05917688873079088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,1536,0.006075555665625467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,128,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,6144,0.035997334453794695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,512,0.012771555946932899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,1536,0.024785778588718835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,2560,0.011716444459226398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,16384,0.04673155479960971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,10240,0.016159110599093966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,16384,0.0236462222205268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,1536,0.005965333431959152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,65536,0.09325155284669663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,10240,0.02776177724202474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,256,0.002513777878549364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,8192,0.027450667487250432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,3072,0.005576000031497743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,768,0.007066667079925537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,2048,0.008368000388145447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,3072,0.007246221933099959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,4096,0.03314133485158285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,8192,0.021232000655598108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,1024,0.004558222161398994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,3584,0.016574222180578444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,12288,0.021241777473025854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,3584,0.009043555292818282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,12288,0.11540444691975911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,3072,0.00553955551650789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,16384,0.0440977778699663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,5120,0.10535555415683323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,16384,0.023634667197863262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,7168,0.1865004433525933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,4096,0.01590488851070404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,10240,0.018200000127156574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,65536,0.12631111674838597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,512,0.019553777244355943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,512,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,512,0.0069857777820693115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,3584,0.017848889032999676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,10240,0.048360887500974864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,10240,0.012130666938092975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,7168,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,256,0.01651822196112739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,2048,0.006568000134494569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,10240,0.02725155486000909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,3072,0.00666133314371109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,2048,0.03559644354714288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,64,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,65536,0.07519466347164579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,256,0.014146667387750415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,7168,0.01276000009642707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,16384,0.03424800104565091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,10240,0.08583733108308579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,512,0.0021928888228204516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,256,0.006607111129495833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,10240,0.014104889498816596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,16384,0.14153156015608045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,3584,0.022312000393867493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,12288,0.024317334095637005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,3584,0.015460444821251763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,2048,0.004520888957712385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,768,0.005736888696750005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,128,0.0021226666867733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,2560,0.011413333316644033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,12288,0.07607022258970472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,12288,0.11627822452121311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,4096,0.008616888688670265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,12288,0.0737066666285197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,128,0.009707555174827576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,12288,0.03189599845144484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,10240,512,0.009868444667922126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,8192,0.01536799967288971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,2560,0.009910222556855943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,768,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,5120,0.011064888702498542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,4096,0.02349688940578037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,10240,0.2111919985877143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,4096,0.00759733302725686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,4096,0.006583111153708563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,12288,0.05925510989295112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,5120,0.02019733356104957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,2560,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,12288,0.016159999701711867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,768,0.012728888955381183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,1536,0.036609778801600136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,64,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,12288,0.029498666524887085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,1536,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,6144,0.016394666499561734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,16384,0.034394668208228216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,512,0.015779554843902588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,1024,0.005659555395444234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,7168,0.008963555925422246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,5120,0.008302222523424361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,128,0.0024968888610601425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,64,0.0018133332745896445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,5120,0.007990222010347579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,10240,0.07080533107121785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,2560,0.010390222072601318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,7168,0.13351821899414062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,10240,0.016232000456915963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,10240,0.01512888901763492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,2048,0.008616888688670265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,256,0.002499555548032125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,6144,0.01921422282854716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,12288,0.03014933400683933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,8192,0.0369777778784434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,768,0.015824889143308003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,1024,0.0032275555034478507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,6144,0.00942666663063897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,768,0.030208888981077407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,1024,0.003564444267087512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,128,0.009001777403884465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,768,0.024751999311976965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,6144,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,6144,0.009311999711725447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,4096,0.02320266597800785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,3072,0.024400000770886738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,3584,0.016113777955373127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,128,0.021346666746669348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,3584,0.027961777316199407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,10240,0.03254133462905884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,7168,0.04680888851483663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,10240,0.018541332748201158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,3584,0.009304000271691216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,256,0.002519999941190084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,2560,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,512,0.006280000011126201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,256,0.01204355557759603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,7168,0.010763555765151978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,2048,0.006593777901596493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,3584,0.0075297777851422625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,12288,0.021060445242457922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,1024,0.0058959999846087555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,5120,0.015855110353893705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,6144,0.01352533366945055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,512,0.02077155477470822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,3072,0.020814221766259935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,768,0.012823111481136747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,4096,0.017186666528383892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,16384,0.025954667064878676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,32,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,3584,0.016514667206340365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,2048,0.006351999938488007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,512,0.005011555635266834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,4096,0.014256889621416727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,256,0.00249777775671747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,2048,0.013798221945762634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,2048,0.009741333623727163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,16384,0.0717866685655382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,2048,0.006004444426960415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,256,0.002141333288616604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,1536,0.00628977765639623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,65536,0.07176444265577528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,768,0.005888888819350137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,6144,0.025288888149791296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,2048,0.006591111007663939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,65536,0.17043644852108428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,5120,0.026320000489552815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,65536,0.1315608951780531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,128,0.006268444574541516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,12288,0.024119110571013555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,65536,0.09469155470530193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,1024,0.006227555374304454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,2048,0.03273066547181871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,10240,0.030609776576360066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,3072,0.016524444023768108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,32,0.0058746664888328975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,8192,0.03303822212749057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,8192,0.020083554916911654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,7168,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,128,0.014095111025704278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,768,0.012776888906955719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,6144,0.010323555933104621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,2048,0.013788444300492605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,65536,0.04859911070929634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,16384,0.015298666225539314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,7168,0.01551555593808492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,12288,0.017887110511461895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,64,0.00665866666369968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,32,0.005909333212508096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,10240,0.06259555286831327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,2048,0.0041884444653987885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,512,0.0028479999552170434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,5120,0.023772444989946153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,12288,0.23160266876220703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,2048,0.006599999964237213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,2560,0.012741333080662621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,6144,0.04001511136690775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,5120,0.006574222197135289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,4096,0.059241778320736356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,65536,0.08306133084826998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,5120,0.01458577811717987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,256,0.00872622181971868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,768,0.006267555471923616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,2048,0.008671999805503422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,12288,0.013891556196742587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,512,0.014498665928840637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,10240,0.0199626667631997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,10240,0.029892444610595703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,5120,0.04909600151909722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,8192,64,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,10240,0.045112000571356885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,1024,0.019098665979173448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,1536,0.00796888851457172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,16384,0.04445599847369724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,6144,0.009023110899660323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,128,0.006298666612969504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,6144,0.058111111323038735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,4096,256,0.03293244375122918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,256,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,1024,0.015156444576051501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,2048,0.006974221931563483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,256,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,12288,0.030198223061031763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,10240,0.015463110473420886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,3072,0.047247999244266085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,1536,0.006280888699822956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,768,0.005913777897755305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,2048,0.008593777815500895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,7168,0.05500177873505486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,65536,0.1488222281138102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,64,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,4096,6144,0.011691555380821228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,32,0.002132444539003902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,2048,0.006589333216349284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,5120,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,65536,0.08441333638297187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,1536,0.003527111063400904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,256,0.020647999313142564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,8192,0.06377955277760823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,3584,0.0107004443804423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,4096,0.008275555239783393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,2048,0.01444533301724328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,10240,0.0229768885506524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,7168,0.009307555854320526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,2048,0.007358222372002072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,6144,0.09492266840404934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,3584,0.3157004515329997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,7168,0.03929155402713352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,3584,0.008004444340864817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,1536,0.0031759999692440033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,10240,0.020818667279349435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,10240,0.011564444336626263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,6144,0.044489777750439115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,6144,0.01295377810796102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,5120,0.012777778009573618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,1536,0.05885777870814005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,512,0.005561777700980504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,1536,0.011740444435013665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,512,0.06330933173497517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,512,0.005859555469618903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,512,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,512,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,6144,0.019291554888089497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,2560,0.08566577566994561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,768,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,7168,0.027428444888856676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,768,0.017196445001496207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,128,0.007260444263617198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,16384,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,8192,0.011029333704047732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,256,0.005197333378924264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,3584,0.007281777759393056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,7168,0.03134311238924662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,5120,0.010042666561073726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,5120,0.008995555341243744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,256,0.0032248888164758682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,12288,0.05115466647677951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,512,0.008007999923494127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,4096,0.08847910828060573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,10240,0.038719111018710665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,7168,0.03024622135692173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,32,0.0021671110557185277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,5120,0.17346666918860543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,65536,0.07755377557542589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,5120,0.013008000122176277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,65536,0.19468711482154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,2048,0.031790223386552595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,8192,0.0181048893266254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,1536,0.07406044668621488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,1536,0.0780257781346639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,256,0.006311111152172089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,2048,0.004223111189073986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,4096,0.006565333240562015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,10240,0.007666666474607255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,5120,0.10895022418763901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,7168,0.020032889313167997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,7168,0.06966311401791044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,16384,0.1557395590676202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,10240,0.0417520006497701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,4096,0.008600888980759515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,4096,0.008020444048775567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,10240,0.013743999931547375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,2560,0.0195448895295461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,768,0.020486222373114694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,3072,0.014107555150985718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,1536,0.008411555654472774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,4096,0.03202222122086419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,3072,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,3584,0.0855555534362793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,6144,0.017842666970358956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,1536,0.003597333199448056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,3584,0.015472888946533203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,64,0.004877333425813251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,256,0.016196444630622864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,6144,0.017112889223628573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,64,0.021350221501456365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,3072,0.006945778098371293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,6144,0.030862222115198772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,1024,0.006179555422729916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,4096,0.009131555755933126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,32,0.0053102220926019884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,10240,0.04300533400641548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,1536,0.004896000027656555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,3072,0.007662222617202335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,3584,0.01978933314482371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,10240,0.012255111502276527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,8192,0.010048888623714447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,1024,0.007939555578761632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,32,0.0018479999982648427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,3072,0.0684773325920105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,3072,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,7168,0.02235200007756551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,8192,0.005538666827811136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,8192,0.01646311084429423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,5120,0.040991110934151545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,2048,0.0041831110914548235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,128,0.028718223174413044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,3072,0.007280000381999546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,4096,0.007279111279381647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,5120,0.017217778497272067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,128,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,5120,0.31493600209554035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,1536,0.025984888275464375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,16384,0.012031111452314587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,64,0.008002666963471307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,8192,0.028148445818159316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,4096,0.007615110940403408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,2560,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,65536,0.059917330741882324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,7168,0.02032355633046892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,10240,0.014789332946141561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,3584,0.03309955530696445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,8192,0.010683555569913654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,65536,0.07369333505630493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,128,0.01407199932469262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,8192,0.014817777607176038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,7168,0.025952888859642878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,256,0.010049777726332346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,2560,0.011766222616036734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,2560,0.02628622286849552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,8192,0.03799022237459818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,1024,0.0034204444123639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,6144,0.01443822185198466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,12288,0.02534755567709605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,64,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,256,0.019163555569118924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,2560,0.004187555362780889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,10240,0.06457777818044026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,1536,0.08326222499211629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,2560,0.018917333748605516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,1024,0.006239999913507038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,64,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,8192,0.011049778097205691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,1024,0.005881777654091517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,768,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,3072,0.006955555743641323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,512,0.003887999802827835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,4096,0.009759110709031424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,2560,0.014022222823566861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,64,0.01585244470172458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,5120,0.004866666677925321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,64,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,12288,0.023451555106374953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,768,0.018026666508780587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,3584,0.011728888584507836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,256,0.016153777639071148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,8192,0.04764266808827718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,128,0.0021955555097924336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,6144,0.12127644485897487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,4096,0.00796622203456031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,8192,0.04749955402480232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,512,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,3584,0.025076443950335186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,512,0.005621333503060871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,256,0.006270222365856171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,128,0.01495822270711263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,8192,0.4631182352701823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,32,0.005514666852023866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,3584,0.007337777978844113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,8192,0.005325333525737126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,10240,0.050901333491007485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,768,0.01699199941423204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,4096,0.010713777608341642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,12288,0.041230221589406334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,128,0.0031635556370019913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,4096,0.017315554949972365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,4096,0.04210844304826525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,7168,0.01369599997997284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,128,0.004960000101062986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,6144,0.01552977826860216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,65536,0.02183377742767334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,16384,0.020308444897333782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,16384,0.015578667322794596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,12288,0.012795555922720166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,3072,0.020978666014141504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,6144,0.009725333087974125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,256,0.01424888935354021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,2560,0.03173244330618117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,768,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,512,0.0144533332851198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,8192,0.013895110951529609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,256,0.016436444388495553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,1024,0.1543511152267456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,7168,0.012055111428101858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,2048,0.00629066675901413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,3072,0.026445334156354267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,32,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,256,0.004277333203289243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,1536,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,65536,0.07910222477383085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,1536,0.005512889060709212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,1536,0.020255999432669747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,512,0.05476444297366672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,8192,0.016920000314712524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,16384,0.043559110826916166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,512,0.017903111047214933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1536,1024,0.003168888803985384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,12288,0.016922665966881644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,1536,0.009302222066455418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,65536,0.20470133092668322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,128,0.004880889008442561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,65536,0.09801777866151597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,3584,0.0336133340994517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,3072,0.013163555827405719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,10240,0.02079288827048408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,3584,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,32,0.011020444333553314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,1024,0.007620444728268518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,2048,0.005724444571468566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,2048,0.006586666736337874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,1536,0.004424889054563311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,4096,0.05400177836418152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,10240,0.012831110921170978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,32,0.004919111314747068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,768,0.003842666745185852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,3584,0.00775022225247489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,512,0.005868444426192178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,512,0.005092444519201915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,4096,0.021631111701329548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,2560,0.02756711178355747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,16384,0.0075431110130416016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,65536,0.10338489214579265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,128,0.025407110651334126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,128,0.006543111056089401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,10240,16384,0.08332622051239014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,65536,0.15446221828460693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,1536,0.015133332875039844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,5120,0.020575111111005146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,16384,0.016883555385801528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,1024,0.052031109730402626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,16384,0.04949422346221077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,12288,0.016399999459584553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,1536,0.010320888625250923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,3072,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,1024,0.061977777216169566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,1536,0.022008889251285132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,6144,0.01956533392270406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,5120,0.026683555708991155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,3072,0.021661332911915247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,12288,0.06302933560477363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,128,0.0021706666383478376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,4096,0.011690666278203329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,1536,0.006238222122192383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,1536,0.006569777925809224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,1536,0.011027555498811932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,2048,0.0058000000814596815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,5120,0.006175111151403851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,12288,5.195935991075304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,2048,0.008004444340864817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,3072,0.009699555734793345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,5120,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,8192,0.0046951112647851305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,7168,0.03144977821244134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,64,0.005592000153329637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,10240,0.05567200316323174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,6144,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,6144,0.022600000103314716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,65536,0.0978142221768697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,2560,0.004862222406599256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,12288,0.025597333908081055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,3072,0.00867377801073922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,3072,0.03479199939303928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,768,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,6144,0.03753866751988729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,5120,0.01163644426398807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,7168,0.011511111425028907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,768,0.003593777616818746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,3072,0.012829333543777466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,6144,0.017864889568752713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,65536,0.245124446021186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,512,0.009017777939637503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,5120,0.023927110764715407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,3584,0.009092444347010719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,32,0.002144888871245914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,3072,0.007256000406212277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,512,0.0028284444577164124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,2048,0.006523555351628198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,6144,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,1536,0.006990222467316522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,8192,0.0192239996459749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,2048,0.01202488856183158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,128,0.0020799999021821553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,7168,0.1854648854997423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,1024,0.008031110796663497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,12288,0.03520888752407498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,1536,0.004537777768241035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,128,0.002171555534005165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,128,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,6144,0.018390221728218924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,4096,0.011530666715568967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,3072,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,768,0.005912888795137405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,6144,0.02065333392884996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,2560,0.0069191112286514705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,6144,0.006591111007663939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,1536,0.014888889259762235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,256,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,512,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,1536,0.0041991112132867174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,1024,0.003846222327815162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,3072,0.0842284427748786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,2048,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,512,0.005962666538026597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,3584,0.018548443913459778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,16384,0.023264888260099623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,7168,0.01645866698688931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,256,0.01578044394652049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,512,0.013364444176355997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,128,0.008921778036488427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,1024,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,128,0.006626666833957036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,1024,0.005610666755172942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,128,0.03148622314135233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,8192,0.012077333198653327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,5120,0.013991110854678683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,12288,0.05047466688685947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,4096,0.027448889282014634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,2560,0.01034044474363327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,2048,0.005498666730191972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,256,0.006609777609507243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,128,0.015429332852363586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,128,0.014098667436175875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,768,0.0035235554807715942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,16384,0.08163822359508939
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,6144,0.02400355537732442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,256,0.014486221803559197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,2560,0.007657777931955125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,4096,0.1848986678653293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,16384,0.2128666639328003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,16384,0.018336888816621568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,3072,0.01831822262869941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,8192,0.014907555447684394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,12288,0.036363555325402155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,10240,1.496212429470486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,3072,0.008674666285514832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,512,0.00830666638082928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,6144,0.029157333903842505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,2048,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,4096,0.00959822204377916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,1024,0.028661333852344092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,3072,0.008647999829716152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,4096,0.009731555150614845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,128,0.014816888504558139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,128,0.005913777897755305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,16384,0.23665687772962782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,1024,0.029436445898479883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,3072,0.01109777804878023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,128,256,0.010016000105275048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,3584,0.19915910561879477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,12288,0.028044445647133723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,512,0.021661332911915247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,8192,0.019176888797018263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,2560,0.021320000290870667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,32,0.0028373334142896864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,5120,0.09795200162463719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,256,0.015255999233987598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,128,0.0021679999513758556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,5120,0.01684533390733931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,64,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,3584,0.025847999585999384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,64,0.002559111143151919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,2560,0.01293333371480306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,7168,0.051303111844592623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,10240,0.021943999661339655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,128,0.014849777022997538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,65536,0.032186667124430336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,10240,0.025370667378107708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,128,0.005690666536490123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,7168,0.025062221619817946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,512,0.02197333342499203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,3072,0.007320000065697565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,512,0.0028044444819291434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,768,0.0028320000403457214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,512,0.05524355504247877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,1536,0.00589155571328269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,1536,0.006269333263238271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,5120,0.006643555644485686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,3072,0.007524444825119442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,256,0.01202400028705597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,4096,0.021306667062971327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,10240,0.025398222936524287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,10240,0.01787644459141625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,1536,0.01824355622132619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,1024,0.051184889343049794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,65536,0.0774551100201077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,2560,0.07986666758855183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,6144,0.01960444450378418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,128,0.011049778097205691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,5120,0.020620443754725985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,256,0.0021297778520319196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,7168,0.013775111072593264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,512,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,4096,0.008300444318188561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,65536,0.08569866418838501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,32,0.0018168888572189545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,256,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,10240,0.013935999737845527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,3072,0.022637334134843614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,2048,0.012716444830099741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,7168,0.11485866705576579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,1024,0.014254222313563028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,10240,0.018181333939234417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,7168,0.02293955617480808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,512,0.003176888864901331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,128,0.014132445057233175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,32,0.005501333210203383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,2048,0.008303999900817871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,3584,0.0329751107427809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,1024,0.007636444436179266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,3584,0.007697777615653143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,65536,0.0683822234471639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,768,0.007289778027269576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,256,0.02544622288809882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,2560,0.023698666029506262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,5120,0.03150399857097202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,768,0.029711110724343195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,6144,1.1939751307169597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,1536,0.005912000106440649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,3072,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,512,0.006612444503439798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,8192,0.04853244291411506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,64,0.004559999952713649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,10240,0.02204622162712945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,768,0.002833777831660377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,768,0.016216889023780823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,768,0.006168889088763132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,5120,0.02367377777894338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,256,0.016209777858522203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,10240,0.024791111548741657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,5120,0.11069333553314209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,4096,0.008286221987671321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,6144,0.041617777612474226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,32,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,3584,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,3584,0.008286221987671321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,768,0.016134222348531086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,768,0.006953777538405524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,65536,0.09494133128060235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,1536,0.006385777973466449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,2560,0.014886221951908536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,65536,0.6303226682874892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,12288,0.0591448876592848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,4096,0.00793688909875022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,6144,0.014800889624489678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,768,0.012485333614879184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,16384,0.046894222497940063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,1536,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,3072,0.07872977521684435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,65536,0.07564266522725423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,12288,0.023372444841596816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,10240,0.012746666868527731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,512,0.004895110925038655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,4096,0.007619555625650618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,10240,0.08801422516504924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,65536,0.34264087677001953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,3584,0.017437333861986797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,512,0.003220444545149803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,16384,0.08172977632946439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,128,0.009940444595283931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,1024,0.020229332976871066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,768,0.003220444545149803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,1024,0.01684711045689053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,3072,0.03085422184732225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,7168,0.030159112479951646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,65536,0.25122043821546763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,1024,0.003481777591837777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,32,0.0025173332542181015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,64,0.0041537777417235905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,16384,0.0270453327231937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,2048,0.015465777781274585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,1024,0.08138933446672228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,10240,0.023004444109068975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,7168,0.02143999934196472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,768,0.0032524443748924467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,3584,0.010438222024175854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,16384,0.125617782274882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,4096,0.011349333657158745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,2048,0.025292444560262892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,12288,0.020920889245139226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,64,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,256,0.005535111245181825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,2560,0.0069288888739215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,8192,0.4677022298177083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,6144,0.013384000294738345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,5120,0.02349511120054457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,12288,0.016399110356966656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,5120,0.22137510776519775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,3072,0.007647111184067196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,4096,0.03320622113015916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,4096,0.01685422162214915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,12288,0.05888622336917453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,10240,0.02160888910293579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,32,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,65536,0.2600746684604221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,4096,0.007300444775157505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,10240,0.034394668208228216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,7168,0.10973688628938462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,6144,0.050223999553256564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,6144,0.009355555805895064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,768,0.014119111829333834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,16384,0.025470222036043804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,5120,0.01481066644191742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,65536,0.5414151085747613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,1024,0.013809777796268463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,1024,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,1024,0.0032213332338465583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,4096,0.0251226673523585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,256,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,3072,0.015536889433860779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,3072,0.0181688881582684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,6144,0.009360888765917884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,3072,0.007278222176763747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,2048,0.02730577852990892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,768,0.02164977788925171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,4096,0.027355555031034682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2048,512,0.004960000101062986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,3584,0.019609777463806998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,128,0.004580444345871608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,1536,0.003565333369705412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,7168,0.03644977675543891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,768,0.005619555711746216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,1536,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,65536,0.4803368780348036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,65536,0.4503733317057292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,8192,0.5193795628017849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,2048,0.008230222596062554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,3584,0.030068443881140813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,4096,0.017788444956143696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,768,0.01993600030740102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,8192,0.02026844521363576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,768,0.0029093333416514923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,768,0.02195466723706987
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,768,0.013419555293189155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,5120,0.010679999987284342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,7168,0.03761511047681173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,1024,0.009017777939637503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,7168,0.053435554107030235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,3584,0.0103511114915212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,7168,0.01662488944000668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,2048,0.028152889675564233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,8192,0.022287999590237934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,3072,0.009537777966923183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,32,0.0022106667359670005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,10240,0.015260444747077094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,65536,1.8242070939805772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,2560,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,256,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,1536,0.007537778053018782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,10240,0.01648977730009291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,3072,0.02773511078622606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,16384,0.14483200179206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,8192,0.03445600138770209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,16384,0.6835289001464844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,768,0.019306666321224637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,32,0.007095110913117726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,4096,0.021373333202468023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,4096,0.24754932191636828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,3584,0.007702222300900354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,8192,0.005517333332035277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,32,0.005335999859703912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,32,0.002036444428894255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,128,0.018608000543382432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,8192,0.03228888909022013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,2048,0.016207999653286405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,65536,0.19751732879214814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3584,12288,0.3766320016649034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,6144,0.009320889082219865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,4096,0.008299555215570662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,5120,0.08059822188483344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,32,0.03122399912940131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,128,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,64,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,2560,0.01621955633163452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,12288,0.0158915552828047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,64,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,3072,0.008295111358165741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,128,0.019695111446910434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,16384,0.01439377831088172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,8192,0.010847110715177325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,3072,0.027031110392676458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,7168,0.09184533357620239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,1536,0.003580444388919406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,6144,0.028022223048739966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,768,0.0038542221817705366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,8192,0.03370399938689338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,16384,0.05080711179309421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,10240,0.0177875558535258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,16384,0.12874933083852133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,65536,0.2386231157514784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,5120,0.02641955514748891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,1536,0.019188443819681805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,3072,0.023319111929999456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,65536,0.2805742157830132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,3584,0.04144088758362664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,7168,0.016920889417330425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,8192,0.012406222522258759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,1024,0.0031848889258172777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,8192,0.01958933307064904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,1536,0.013757333159446716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,10240,0.008993777963850234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,10240,0.01202488856183158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,32,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,3584,0.024792888098292883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,16384,0.01927199959754944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,10240,0.035691556003358625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,16384,0.015133332875039844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,6144,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,16384,0.046945777204301625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,1024,0.015461333923869662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,1536,0.014491554763582019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,128,0.008995555341243744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,256,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,4096,0.022845332821210224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,10240,0.012387555506494312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,1536,0.014469333820872836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,1536,0.006239999913507038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,1024,0.008954666554927826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,7168,0.011698666546079846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,4096,0.02700533303949568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,16384,0.023818666736284893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,1536,0.00981155534585317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,768,0.016190222567982145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,2560,0.01036088913679123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,32,0.002120888895458645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,2560,0.007919111185603673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,5120,0.016899555921554565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,256,0.005209777918126848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,768,0.015797333584891427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,768,0.008227555288208855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,2048,0.012734221915404001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,3584,0.015464888678656684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,3584,0.007617777420414819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,3072,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,3584,0.21406133969624838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,3584,0.0048942222363419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,8192,0.014012444350454541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,3072,0.01587199999226464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,128,0.023931556277804907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,256,0.00790666706032223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,12288,0.012839111189047495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,4096,0.01587199999226464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,65536,3.9306810167100696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,1536,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,3072,0.01148000028398302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,512,0.013071111506885953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,3584,0.016361777981122334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,5120,0.01817244456874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,256,0.0052328887912962176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,256,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,10240,0.028526223368114893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,1024,0.009342222577995723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,12288,0.05333066648907132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,16384,0.033626665671666466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,16384,0.03424800104565091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,6144,0.0094373333785269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,7168,0.045124444696638316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,256,0.007693332930405934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,1536,0.005991999887757831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,65536,0.1684568855497572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,10240,0.012289777398109436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,3584,0.01555377741654714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,4096,0.016879111528396606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,16384,0.033406221204333834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,4096,0.04205866654713949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,512,0.0158924443854226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,3072,0.007315555380450354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,16384,0.04471822248564827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,512,0.0032640000184377036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,16384,0.059452447626325816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,2560,0.009104000197516548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,10240,0.011912888950771756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,2048,0.04188088907135857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,5120,0.01660355594423082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,7168,0.044194665220048696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,1024,0.006587555425034628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,2048,0.009029332962301042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,128,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,4096,0.018910222583346896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,12288,0.015103111664454142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,7168,0.02491377790768941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,8192,0.010064000056849586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,512,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,4096,0.03291733397377862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,768,0.03904177745183309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,128,0.007633777956167857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,512,0.012431999875439538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,512,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,512,0.011032000184059143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,1536,0.013165333204799227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,16384,3.4200880262586804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,4096,0.010838222172525195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,10240,0.04471733172734579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,5120,0.0410186681482527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,32,0.006561777657932705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,768,0.005255110975768831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,4096,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,7168,0.013263111313184103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,2560,0.02051999999417199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,2560,0.018601778480741713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,8192,0.010182222558392419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,2560,0.02443466583887736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,2048,0.007548444800906711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,12288,0.04643911123275757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,5120,0.00832266691658232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,10240,0.2007519933912489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,1536,0.009930666950013902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,512,0.0026364444444576898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,10240,0.04247911108864678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,512,0.017909333109855652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,1536,0.00630044440428416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,256,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,65536,3584,0.028224001328150432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,10240,0.1907555527157254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,16384,0.45975732803344727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,1536,0.011379555695586734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,4096,0.02162844439347585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,128,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,5120,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,12288,0.019504888190163504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,2048,0.004202666795916027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,8192,0.08777066734102036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,12288,0.086444444126553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,3072,0.028265777561399672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,10240,0.03466044531928168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,12288,0.014145778285132514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,7168,0.04769777920511034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,1024,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,1536,0.007990222010347579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,2560,0.016241777274343703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,10240,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,3584,0.011045333411958484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,4096,0.008015111088752747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,3584,0.08796266714731853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,2560,0.011310222248236338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,1536,0.03292355603641934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,768,0.00554666668176651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,7168,0.026737777723206416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,1536,0.00619911112719112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,65536,0.1898124482896593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,128,0.014222222897741528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,65536,0.09396533171335857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,1536,0.018590221802393597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,65536,0.07745777898364596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,768,0.006837333242098491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,3072,0.008387555678685507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,512,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,6144,0.04278399878078037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,3072,0.011022222538789114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,3072,0.0976497795846727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,4096,0.04054044352637397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,512,0.013795555465751223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,7168,0.11634933948516846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,512,0.005532444351249271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,10240,0.012206222448084088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,256,0.014311111635631986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,2048,0.004664888812435998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,512,0.00598933340774642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,8192,0.012727999852763282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,256,0.004946666873163647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,4096,0.01591466698381636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,1536,0.010713777608341642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,768,0.0029164445069101122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,768,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,3584,0.02363555630048116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,4096,0.01755288905567593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,4096,0.01721600029203627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,12288,0.016527111331621807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,512,0.0025395556456512874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,12288,0.04928266670968798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,1024,0.00907199995385276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,5120,0.008419555922349295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,4096,0.004208000169859992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,12288,0.018882667024930317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,6144,0.008986666798591614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,8192,0.026816000541051228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,768,0.016262221667501662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,3072,0.023747555083698694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,2048,0.008336000144481659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,7168,0.009720000127951304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,3072,0.012065778175989786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,128,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,8192,0.01611911091539595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,3584,0.007289778027269576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,6144,0.009326222042242685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,3584,0.02299822204642826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,16384,0.024064888556798298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,5120,0.05794399976730347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,768,0.002514666567246119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,1536,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,1024,0.014797333214018079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,3072,0.02762311034732395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,10240,0.3833297888437907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,1024,0.004252444538805219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,512,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,65536,0.2774346669514974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,2560,0.006240889016124938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,2048,0.009082666701740688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,128,0.014448889427714877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,5120,0.01038044442733129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,12288,0.023756444454193115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,3584,0.027976890405019123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,1024,0.005984000033802456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,5120,0.008575999902354347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,6144,0.032348444064458214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,65536,0.07041777504814996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,512,0.004948444250557157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,1024,0.00564177789621883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,2048,0.03994133406215244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,64,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,3584,0.05368177758322822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,256,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,8192,0.028386665715111628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,12288,0.10251110792160034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,2560,0.023235556152131822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,6144,0.02129511038462321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,4096,0.011100444528791638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,512,0.00870222184393141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,10240,0.064410666624705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,1024,0.007303111255168915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,128,0.006414222220579783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,10240,0.012063110868136087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,768,0.002801777794957161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,512,0.0029360000044107437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,5120,0.005621333503060871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,12288,0.03644888930850559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,256,0.0022560000005695554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,1024,0.013408888545301227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,5120,0.015388444066047668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,16384,0.046041776736577354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,16384,0.031197332673602637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,768,0.005953777581453323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,4096,0.013774221969975365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,4096,0.02772444486618042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,1024,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,3072,0.014455111490355598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,5120,0.19071377648247612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,8192,0.04370222157902188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,16384,0.017998221847746108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,3072,0.009503111243247986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,2048,0.00865955568022198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,4096,0.027223111854659185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,2048,0.0073902226156658595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,7168,0.009840000006887648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,64,0.043939554029040866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,1536,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,65536,2.6918765174018016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,768,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,768,0.006298666612969504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,16384,0.0218231115076277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,1024,0.00795999997191959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,1536,0.007355555891990662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,2560,0.009103111094898647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,7168,0.0470728874206543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,2560,0.007994666695594788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,2560,0.15525510576036242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,3072,0.007643555601437886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,65536,0.12336977322896321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,16384,0.04012088974316915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,12288,0.019302222463819716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,1024,0.013765333427323235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,7168,0.030609776576360066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,6144,0.03460977805985345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,5120,0.016354666815863717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,128,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,2560,0.043471111191643604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,10240,0.04583288894759285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,12288,0.013006221916940479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,10240,0.023342221975326538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,1024,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,5120,0.01311822235584259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,6144,0.03594400154219733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,4096,0.01832088828086853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,1536,0.006303999986913469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,10240,0.03324266605907016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,65536,0.07753422525193956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,6144,0.011891555454995898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,65536,0.055854221185048424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,3584,0.005919999960396025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,1024,0.00590133335855272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,5120,0.03001777662171258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,6144,0.2254862255520291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,16384,0.02497422198454539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,5120,0.03093155556254917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,256,0.035721778869628906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,1024,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,6144,0.019168888529141743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,16384,0.04190577732192146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,6144,0.013806222213639153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,6144,0.018245332770877414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,5120,0.05086133215162489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,10240,0.030985779232449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,3584,0.008361777497662438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,6144,0.040050665537516274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,2048,0.009372444616423713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,256,0.002826666666401757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,3584,0.1354400051964654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,5120,0.0481644438372718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,64,0.005626666463083691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,5120,0.009709333380063375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,4096,0.010204444328943888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,32,0.0021457777669032416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,5120,0.008618666893906063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,2560,0.018608889646000333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,7168,0.018627555833922494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,8192,0.2670293384128147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,6144,0.014596444037225513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,64,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,65536,0.16229511631859675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,4096,0.01956622302532196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,512,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,16384,0.021588444709777832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,16384,0.05138755506939358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,6144,0.004644444419278039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,5120,0.007371555599901412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,1024,0.008045333127180735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,7168,0.02277244461907281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,7168,0.010023111270533668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,3584,0.007576888634098901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,5120,0.0083155557513237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,2560,0.007259555160999298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,8192,0.01657777859105004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,16384,0.01883288886811998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,768,0.09209511015150283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,3584,0.007133333219422235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,3072,0.011377777490350934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,768,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,256,0.002149333349532551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,10240,0.05459288756052653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,10240,0.08335200283262466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,512,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,16384,0.029356443219714697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,512,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,256,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,5120,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,6144,0.034127109580569796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,12288,0.08520978026919895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,128,0.0063324446479479475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,4096,0.004467555632193883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,1024,0.01887822151184082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,5120,0.014127999544143677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,7168,0.025565332836574976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,8192,0.016992888516849942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,768,0.0034933334423436057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,64,0.006597333484225803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,128,0.010433778166770935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,128,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,12288,0.12933066156175402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,1536,0.007312888900438945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,128,0.0028524444335036804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,7168,0.010026666853162978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,4096,0.011323555476135679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,1024,0.0042133331298828125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,256,0.016154666741689045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,65536,0.07536177502738105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,8192,0.01184800018866857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,8192,0.02583377725548214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,5120,0.005382222019963794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,10240,0.06352533234490289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,12288,0.17618221706814238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,768,0.017743999759356182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,32,0.00563466673096021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,3584,0.029265777932273015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,12288,0.04671555426385668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,1024,0.0032008888406885993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,32,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,128,0.003536000019974179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,12288,0.03411644366052415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,65536,3.8247172037760415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,128,0.0028497777465316984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,8192,0.008692444198661381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,1536,0.013782222237851886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,32,0.003531555748648114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,5120,0.012783110969596438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,768,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,8192,0.02032177812523312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,7168,0.02680977847841051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,6144,0.005195555587609609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,6144,0.0199644449684355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,2560,0.010077333284748925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,64,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,768,0.005568888866239124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,2048,0.0063777777055899305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,1536,0.023615999354256526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,1536,0.02130577796035343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,3584,0.005617777920431561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,8192,0.04102133380042182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,64,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,768,0.01684533390733931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,3584,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,32,0.006608888920810487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,1024,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,5120,0.005588444570700328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,6144,0.02070755594306522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,6144,0.015234667393896313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,4096,0.00774133370982276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,12288,0.014800889624489678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,1024,0.0032426667296224167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,3072,0.015838222371207345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,4096,0.03882755504714118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,5120,0.08146044280793932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,3072,0.03669333457946777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,3584,0.004564444637960858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,65536,0.14981867207421198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,7168,0.016193777322769165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,768,0.017454221844673157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,2048,0.025256888733969793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,7168,0.030219554901123047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,128,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,32,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,2048,0.004597333156400257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,2560,0.0654053356912401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,5120,0.01219466659757826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,1536,0.019886222150590684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,256,0.002509333400262727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,12288,0.02069422271516588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,8192,0.016948444975747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,4096,0.022828444838523865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,1024,0.006960888703664144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,32,0.007607999775144789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,2048,0.006240889016124938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,5120,0.008227555288208855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,3072,0.023046221998002794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,6144,0.01943555474281311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,6144,0.00810222244924969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,4096,0.008741333252853816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,6144,0.00904177791542477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,6144,0.02623733381430308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,16384,0.018056000272432964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,1024,0.006613333192136552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,2048,0.012894222305880653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,5120,0.022477333744366963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,512,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,3584,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,2560,0.009071110851234859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,10240,0.028312000963422988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,32,0.0028355556229750314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,768,0.017518222332000732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,7168,0.010664000279373594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,16384,0.02626577847533756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,1024,0.005598222215970357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,16384,0.032913777563307024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,10240,0.010122666756312052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,12288,0.18657155831654867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,6144,0.13681422339545354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,65536,0.009655999640623728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,128,0.015800888339678448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,6144,0.009030222064918941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,1536,0.023319111929999456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,512,0.002132444539003902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,2048,0.018231110440360177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,16384,0.025705777936511572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,2048,0.010982222027248807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,1536,0.007298666569921706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,256,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,3072,0.01939377850956387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,65536,0.08968355258305867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,6144,0.011146666275130378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,512,0.00833777752187517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,8192,0.01034400032626258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,3584,0.020805334051450092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,65536,0.060989333523644336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,128,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,12288,0.02123111155298021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,6144,0.012328888807031842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,5120,0.012050666742854647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,10240,0.018213333355055917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,2048,0.01864355636967553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,512,0.012510221865442065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,2560,0.009649777577983009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,10240,0.05141599973042806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,7168,0.01276000009642707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,10240,0.016231111354298063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,128,0.004920888692140579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,1536,0.019741333193249173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,4096,0.03584355447027419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,1536,0.00370133337047365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,6144,0.02609777781698439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,4096,0.19269955158233643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,128,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,2560,0.07950755622651842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,768,0.0038444445365005066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,3072,0.006423111177153057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,512,0.006299555715587404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,3584,0.015814221567577787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,512,0.0029315555261241067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,3072,0.007631999750932057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,10240,0.042091555065578885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,3584,0.005975111077229182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,512,0.011872000164455838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,4096,0.008972444468074376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,3072,0.026925333672099646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,512,0.01146488885084788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,128,0.0052639999323421055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,10240,6.841454399956597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,4096,0.08664621909459432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,3072,0.011072888970375061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,512,0.021003555920388963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,2048,0.017248888810475666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,128,0.018910222583346896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,3072,0.01626044511795044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,8192,0.013415999710559845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,7168,0.08123377958933513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,3072,0.009394666386975182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,10240,0.01755644381046295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,3072,0.08236800299750434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,8192,0.9408213297526041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,256,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,10240,0.08954488568835789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,5120,0.3611111111111111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,2560,0.019609777463806998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,2560,0.00905511114332411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,1024,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,2560,0.009347555538018545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,12288,0.03498133354716831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,1536,0.0037600000699361167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,256,0.008799110849698385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,65536,0.15363289250267878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,12288,0.039055999782350324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,1024,0.02257777750492096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,12288,0.48929511176215273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,6144,0.04996977912055122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,1024,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,512,0.05925155348247952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,32,0.003567111161020067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,10240,0.018060444129837882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,3584,0.008015111088752747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,512,0.01382844481203291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,256,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,3584,0.014107555150985718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,4096,0.016864889197879367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,12288,0.05466577741834852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,65536,0.07485689057244195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,10240,0.019938665959570143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,5120,0.008671999805503422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,16384,0.023800889650980633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,1536,0.012772444221708508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,32,0.008622222476535374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,10240,0.013045333325862885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,2560,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3584,16384,0.03446755475468106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,1536,0.00628266649113761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,4096,0.03123733401298523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,1536,0.0035137778355015647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,4096,0.009930666950013902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,32,0.0018942221585247253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,10240,0.017924444542990792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,6144,0.02535733415020837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,3584,0.028235554695129395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,3584,0.026657778355810378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,512,0.006196444647179709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,512,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,128,0.0028133332315418455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,1536,0.006274666637182236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,512,0.0028613333900769553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,5120,0.04895022180345324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,6144,0.009373333719041612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,5120,0.0136808885468377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,8192,0.010407999985747868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,16384,0.05285155442025927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,1536,0.00591555568906996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,2560,0.029519110918045044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,128,0.005561777700980504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,2560,0.004914666629499859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,1024,0.036659555302725896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,4096,0.011157333023018308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,4096,0.026737777723206416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,4096,0.011391110718250275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,3584,0.008309333688682979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,768,0.019578667150603402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,8192,0.01533600025706821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,10240,0.026854222019513447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,1536,0.014194667339324951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,32,0.061434666315714516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,1536,0.02200711104604933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,6144,0.060469335979885526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,16384,0.019544000426928203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,768,0.005505777895450592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,64,0.06003377834955851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,10240,0.040016889572143555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,4096,0.00925333301226298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,3072,0.006587555425034628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,7168,0.03432088759210374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,3584,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,2048,0.006682666639486949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,4096,0.013624889155228933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,768,0.017587555779351127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,5120,0.012488889197508493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,16384,0.03017688790957133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,5120,0.03331289026472304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,12288,0.03767199979888068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,7168,0.01852799952030182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,3584,0.007913777397738563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,6144,0.03797511259714762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,5120,0.004247111164861255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,6144,0.010130667024188572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,1024,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,65536,0.05703733364741007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,4096,0.011642666326628791
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,32,0.0019191110299693213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,2560,0.01204355557759603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,6144,0.036743998527526855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,6144,0.03397599856058756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,7168,0.281713777118259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,2048,0.01091733326514562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,8192,0.021631111701329548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,3584,0.03081688947147793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,3072,0.007439110842016008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,256,0.00490133340160052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,2048,0.011422221859296164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,768,0.01479200025399526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,7168,0.0344471103615231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,65536,0.09271022346284653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,512,0.016586666305859882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,512,0.008622222476535374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,1536,0.0041440000964535605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,256,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,16384,2.635345882839627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,768,0.00553688903649648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,32,0.005252444495757421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,512,0.019924443629052904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,4096,0.007243555453088548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,6144,0.04316622349951002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,768,0.015834665960735746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,8192,0.015423110789722867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,4096,0.02323644525474972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,12288,0.01754666699303521
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,2560,0.007239110767841339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,8192,0.014831999937693277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,10240,0.07275733020570543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,6144,0.010688888529936472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,1024,0.003185777821474605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,12288,0.06876533561282687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,65536,0.3359564410315619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,1024,0.006612444503439798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,256,0.014477334088749357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1024,64,0.005587555468082428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,8192,0.022240888741281297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,64,0.006600889066855113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,16384,0.026905778381559584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,256,0.004591111093759537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,32,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,6144,0.0849271085527208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,2048,0.007300444775157505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,5120,0.11092888646655613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,128,0.010240889257854885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,512,0.04071288969781663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,3072,0.06167288621266683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,8192,0.05052444338798523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,128,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,64,0.005897777775923411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,10240,0.02331111166212294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,8192,0.015527110960748462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,64,0.0025235555238193935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,8192,0.06711110803816053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,5120,0.012455110748608908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,10240,0.0456853343380822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,256,0.008473777936564552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,1536,0.0063884444534778595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,768,0.005607999861240387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,1024,0.008752000000741746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,128,0.004940444396601783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,5120,0.013425778183672162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,6144,0.007983999947706858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,3072,0.007991111112965478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,4096,0.005884444548024072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,12288,0.022634666826989915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,1024,0.011415999796655444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,3072,0.007318221860461765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,3072,0.023058666123284235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,8192,0.01386666629049513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,4096,0.0074755557709270054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,6144,10240,0.032687998480266996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,128,0.0037333332002162933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,6144,0.009919111099508073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,12288,0.018903111418088276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,3584,0.008904000123341879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,65536,0.0910702215300666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,768,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,10240,0.04250133368704054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,3072,0.6536462042066786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,3584,0.015489778584904142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,12288,0.01144444445768992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,8192,0.010994666980372535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,768,0.006218666831652324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,1536,0.005877333382765452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,8192,0.0440684457619985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,3072,0.061287111706203885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,3072,0.008052444292439355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,12288,0.035599999957614474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,1536,0.004206222378545337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,768,0.006656000183688269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,512,0.007735999921957652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,512,0.002479111154874166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,512,0.01447111037042406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,8192,0.3472417725457086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,2048,0.006239111224810283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,5120,0.018722666634453666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,65536,0.0665608909395006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,128,0.011064888702498542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,8192,0.026910222238964502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,65536,0.2064124478234185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,1536,0.010413332945770687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,64,0.0021333334346612296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,4096,0.023123555713229712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,1024,0.006009777800904379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,1536,0.004190222256713443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,8192,0.1182755496766832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,65536,0.07175911135143705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,10240,0.026720000637902155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,16384,0.05412800113360087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,65536,0.10341155529022217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,128,0.01515111161602868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,2560,0.032631112469567194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,10240,0.008830221990744272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,2048,0.0047315553658538395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,7168,0.010370666782061258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,3584,0.0536008874575297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,16384,0.009752000371615091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,6144,0.020933333370420668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,6144,0.028536889288160536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,256,0.005746666755941179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,5120,0.012598222328556908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,512,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,10240,0.028904000918070476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,3072,0.03209422363175286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,1536,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,256,0.014183110660976835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,4096,0.009441778063774109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,3072,0.037351999017927386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,2048,0.017217778497272067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,2048,0.006586666736337874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,1536,0.012129777835475074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,16384,0.017293334007263184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,512,0.005629333357016246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,512,0.01313511116637124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,2048,0.026735110415352717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,7168,0.03073422114054362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,256,0.009032888544930352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,8192,0.11372533109452988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,64,0.007958222594526079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,2048,0.019489778412712943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,4096,0.004389333228270213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,4096,0.008904889225959778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,3072,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,1024,0.004547555413511065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,128,0.005800888770156437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,5120,0.027152889304690894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,2560,0.027816000911924574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,768,0.005281777845488654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,7168,0.020374221934212577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,2560,0.015744000673294067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,12288,0.01853155593077342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,6144,0.015170666906568738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,10240,0.015135111080275642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,12288,0.19693599806891549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,1536,0.006601777755551868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,128,0.011396444506115384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,1536,0.01754577789041731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,768,0.019654222660594516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,1536,0.006614222294754452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,16384,0.022950222094853718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,1536,0.018539556198649935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,6144,0.010377777947319878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,7168,0.03803999887572394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,2560,0.006918222126033571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,16384,0.057373331652747266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,16384,0.026776888304286536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,128,0.005967111223273807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,128,0.01516177753607432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,16384,0.02881866693496704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,16384,0.006237333433495627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,65536,0.21901688310835096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,2048,0.013794666363133324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,5120,0.03066133459409078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,1536,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,5120,0.030842665168974135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,4096,0.007661333514584436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,7168,0.02233422299226125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,768,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,12288,0.10230311420228745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,4096,0.007613333563009898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,65536,0.11000800132751465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,8192,0.017609778377744887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,10240,0.016034666034910414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,128,0.013812444276279874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,2560,0.02265600032276577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,128,0.014094221923086377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,5120,0.00979022185007731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,768,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,4096,0.009320889082219865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,64,0.006562666760550604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,3072,0.007608888877762689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,1024,0.004540444248252445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,1536,0.011384888655609555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,2048,0.044779555665122144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,65536,0.1407511101828681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,16384,0.017958222164048087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,1024,0.0028924445311228433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,1024,0.011047999891969891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,3584,0.018505778577592637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,512,0.0024702221983008915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,1536,0.022287999590237934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,512,0.013011555704805585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,768,0.022967111733224656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,4096,0.01442577772670322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,512,0.005041777673694822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,7168,0.05827199750476413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,65536,0.061193777455223926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,3072,0.011076444553004371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,32,0.00425155543618732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,7168,0.07644710938135783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,2560,0.015124445160230001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,2048,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,4096,0.025240000751283433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,3584,0.007664889097213745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,256,0.0058257778485616045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,8192,0.023359111613697473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,16384,0.02496533261405097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,2560,0.014299554957283868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,1024,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,12288,0.0439662237962087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,12288,0.02052800026204851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,6144,0.02825777729352315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,5120,0.14739910761515299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,3584,0.01034577770365609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,256,0.007308444215191736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,2560,0.029599110285441082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,2560,0.013394667042626275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,3072,0.010250666903124915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,5120,0.010422222316265106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,512,0.01665866706106398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,5120,0.038683556848102145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,2560,0.01354311075475481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,256,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,4096,0.007357333269384172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,5120,0.056418663925594754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,12288,0.06104711029264662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,1536,0.023393778337372675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,32,0.002521777732504739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,16384,0.12134044700198704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,3584,0.03423555692036947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,1024,0.009604444106419882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,12288,0.008565333154466417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,32,0.0069022224181228215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,10240,0.005912000106440649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,256,0.004857777721352047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,3072,0.0069422221018208405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,512,0.01956355571746826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,768,0.0028488888508743713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,768,0.0070266665683852295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,768,0.002805333377586471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,128,0.012827555338541666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,128,65536,0.07152799765268962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,5120,0.06195022000206841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,768,0.005623111294375525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,128,0.013800000150998434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,3072,0.08941600057813857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,256,0.0038693332009845306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,1536,0.006958222223652734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,3072,0.13439555962880453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,1536,0.0362755556901296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,3584,0.028449777099821303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,32,0.006143110907740063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,12288,0.05322755707634819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,2048,0.02075111038155026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,3072,0.007378666765160031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,6144,0.008952000074916417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,2048,0.03398133317629496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,1024,0.003206222214632564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,32,0.005535999933878581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,65536,0.1306933297051324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,65536,0.055777774916754834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,32,0.00656444455186526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,65536,0.14085777600606283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,3584,0.018603555030292936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,5120,0.008310221963458592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,32,0.002151111140847206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,2048,0.006596444381607904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,4096,0.013923555612564087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,7168,0.07567289140489367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,5120,0.014262222581439547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,1536,0.00665511108107037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,3072,0.013726222018400827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,16384,0.04175466630193922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,768,0.0029973333908451926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,768,0.00741511086622874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,256,0.017122666041056316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,256,0.02993333339691162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,2048,0.007280888656775157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,7168,0.014464888307783337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,65536,0.198235551516215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,3584,0.007963555554548899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,2048,0.005943111247486538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3072,256,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,6144,0.014005333185195923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,8192,0.012871111432711283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,1024,0.014167111780908374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,3584,0.054551111327277295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,1536,0.003565333369705412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,64,0.005979555762476391
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,6144,0.012784889174832238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,3072,0.013166222307417126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,3072,0.015354666444990369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,3072,0.008014221986134848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,2048,0.010834666589895884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,12288,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,8192,0.018350222044520907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,12288,0.018931556079122756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,7168,0.042507555749681264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,512,0.012891555825869242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,512,0.005635555419656966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,4096,2048,0.025087111526065405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,12288,0.013983110586802164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,1536,0.0058604445722368025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,65536,0.02385777731736501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,2560,0.01575377749072181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,3072,0.005208000126812193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,2560,0.004926222066084544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,3584,0.004261333495378494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,256,0.011358222199810876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,7168,0.017214222086800467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,128,0.03364888827006022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,65536,1.5635058085123699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,1024,0.01685688893000285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,7168,1.010576883951823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,10240,0.009150221943855286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,8192,0.007271111011505127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,65536,5.6765085856119795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,64,0.00250133333934678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,10240,0.017830222845077515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,32,0.003168888803985384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,256,0.02331999937693278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,1536,0.009348444640636444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,6144,0.21468444665273032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,4096,0.010259555445777046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,65536,0.8383244408501519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,2048,0.02909066610866123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,6144,0.017920000685585868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,32,0.013455111119482251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,8192,0.02076088885466258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,3072,0.008822222550710043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,1024,0.006203555398517185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,3584,0.013843555417325763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,4096,0.007999999655617608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,32,0.004576888763242298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,5120,0.015491555134455362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,32,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,2048,0.008357333640257517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,16384,0.030543999539481267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,768,0.00906755526860555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,65536,0.02642400066057841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,5120,0.008456000023418004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,2560,0.006913777854707506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,2560,0.024007999234729346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,256,0.005596444424655702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,16384,0.044304887453715004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,1024,0.00566133318675889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,128,0.0038853333228164246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,6144,0.020441777176327176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,3584,0.025183111429214478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,16384,0.3661697705586751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,32,0.001869333287080129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,12288,0.7753689024183484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,4096,0.011183111204041375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,3072,0.024526221884621516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,768,0.005535111245181825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,4096,0.022611555125978258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,768,0.00387999994887246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,2048,0.15702933735317653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,256,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,4096,0.010455999937322406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,768,0.005872888697518243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,1536,0.006031111296680238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,3584,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,12288,0.059656891557905406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,12288,0.03554577628771464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,3584,0.004875555634498596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,256,0.014874666929244995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,2560,0.01275466630856196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,1536,0.006043555421961679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,512,0.01518133282661438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,16384,0.08111200067732069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,128,0.002132444539003902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,8192,0.1667333311504788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,256,0.016855110724767048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,2560,0.01277599980433782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,16384,0.09845155477523804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,10240,0.013392888837390475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,1536,0.031790223386552595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,1536,0.006216889040337668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,768,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,10240,0.15117155181037054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,7168,0.010712888505723743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,65536,0.015728889240158927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,3072,0.020224000016848247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,65536,0.05337688989109463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,3584,0.025576000412305195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,3584,0.007616000043021307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,3584,0.02236888971593645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,32,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,5120,0.03883644607332017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,768,0.0028764444092909494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,3072,0.0765848888291253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,1024,0.005577777822812398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,2560,0.009336888790130615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,768,0.020619556307792664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,12288,0.019871110717455547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,1536,0.015896888242827523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,256,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,768,0.005973333285914526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,1536,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,64,0.011520889070298938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,6144,0.010059555371602377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,1024,0.004461333155632019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,768,0.007015110717879401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,3584,0.39025243123372394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,8192,0.010441777606805166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,2560,0.006919999917348226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,3072,0.03919822308752272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,512,0.005518222020732031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,12288,0.013808888693650564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,64,0.03233511249224345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,4096,0.005738666488064661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,3072,0.09238755702972412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,10240,0.011731555892361535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,16384,0.00812533332241906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,32,0.0058666666348775225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,16384,0.022837332553333704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,5120,0.029782222376929388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,65536,0.15424533685048422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,2560,0.006640888750553131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,2560,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,128,0.021969777014520433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,7168,0.015981333123313058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,1024,0.017845332622528076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,8192,0.04106400079197354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,5120,0.03344800074895223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,64,0.0029271110478374693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,7168,0.013895110951529609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,16384,0.21491644117567274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,2560,0.02276533345381419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,128,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,1536,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,12288,0.02030755579471588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,768,0.03317155440648397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,256,0.01000711073478063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,65536,0.48848888609144425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,32,0.0021662221600612006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,12288,0.048933333820766874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,2560,0.026399999856948853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,2560,0.006603555546866522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,6144,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,2560,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,1024,0.006616000086069107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,128,0.014440000057220459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,3584,0.007511999871995714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,32,0.0018613334331247543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,512,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,6144,0.13127466042836508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,2048,0.060312887032826744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,65536,0.10778133074442546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,512,0.012789333032237159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,128,0.010354667074150508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,256,0.004960888789759742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,128,0.012060444388124677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,2048,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,2048,0.004237333519591225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,5120,0.023750222391552393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,256,0.013250666360060373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,512,0.004911111046870549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,768,0.00319377767542998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,3584,0.01386755539311303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,12288,0.023120888405376013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,4096,0.09673511319690281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,2048,0.013144888811641268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,1024,0.00328177772462368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,768,0.013456000222100152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,7168,0.02864266766442193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,128,0.01181244436237547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,128,0.0022657778528001574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,256,0.004889777965015835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,3584,0.03899911046028137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,256,0.005712000032265981
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,2560,0.029401779174804688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,768,0.003944888710975647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,1024,0.013777777552604675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,2048,0.006639110959238476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,2048,0.006252444452709622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,768,0.005444444302055571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,6144,0.01445777714252472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,768,0.015462223026487561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,256,0.0052684446175893145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,128,0.004914666629499859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,10240,0.30555110507541233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,6144,0.02370577719476488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,128,0.014103111293580798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,10240,0.004659555438492033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,1024,0.015834665960735746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,1024,0.005838222387764189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,3072,0.018602665927675035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,3584,0.027495111028353374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,12288,0.03423110975159539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,768,0.007263999846246507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,5120,0.047594666481018066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,3584,0.005956444475385878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,4096,0.007628444168302748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,12288,0.04053511222203573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,256,0.005091555416584015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,3584,0.01000088867213991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,768,0.008756444685988957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,65536,0.07546755340364245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,1536,0.10637244251039292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,65536,0.1208488941192627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,12288,0.01960177719593048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,32,0.0018577778504954444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,64,0.00592533333433999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,2560,0.01479022204875946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,1024,0.005601777798599667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,7168,0.009675555759006077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,65536,0.41677067014906144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,1024,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,3072,0.017059556312031217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,512,0.00665866666369968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,1536,0.014473777678277759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,32,0.006203555398517185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,2560,0.012327999704413943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,7168,0.012051555845472546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,4096,0.016910221841600206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,10240,0.04131644301944309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,7168,0.08780622482299805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,2560,0.02331111166212294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,64,0.005543111099137201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,65536,0.8831626574198405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,4096,0.007988444632954067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,6144,0.04204444421662224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,64,0.002480888946188821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,3584,0.011390222443474663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,12288,0.04733688963784111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,32,0.010378667049937779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,7168,0.010004444254769219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,12288,0.023090667194790308
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,7168,0.0188755558596717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,8192,0.01622577839427524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,128,0.0021315556433465746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,1536,0.01826400061448415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,10240,0.02292266653643714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,1024,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,1024,0.019167110323905945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,2048,0.005568000177542369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,128,0.02151022189193302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,32,0.008059555457697975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,5120,0.01888533267709944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,10240,0.025619554850790236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,5120,0.011926222178671094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,8192,0.02199911077817281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,8192,0.019223110543357003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,128,0.004916444420814514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,256,0.17172888914744058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,3072,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,2048,0.007128000259399414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,2560,0.007583111524581909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,768,0.010808889236715106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,1024,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,16384,0.040228443013297185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,4096,0.009363555245929295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,5120,0.3355875545077854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,3072,0.028824888997607764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,32,0.004080888711743885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,16384,0.025048888391918604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,6144,0.030053334103690252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,7168,0.01959288948112064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,65536,0.07082755698098077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,5120,0.008854221966531541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,2560,0.004905777672926585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,1536,0.015321777926550971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,2560,0.015094222293959724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,128,0.011345778074529437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,3072,0.0069395556218094295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,12288,0.02837688724199931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,8192,0.03190133306715224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,6144,0.008964444200197855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,2560,0.04797600044144524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,7168,0.009499555660618676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,1024,0.004874666531880696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,256,0.007260444263617198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,128,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,2048,0.0041697778635554844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,64,0.004933333231343163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,64,0.0017866666118303935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,3072,0.1887777778837416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,10240,0.017458667357762653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,3072,0.013779555757840475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,128,0.005239999956554837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,16384,0.07397155629263984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,6144,0.01035822182893753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,1536,0.006582222051090664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,6144,0.016944888565275405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,256,0.014448889427714877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,256,0.014102222190962898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,8192,0.01036622209681405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,4096,0.013739555246300168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,64,0.001793777777089013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,3072,0.01779111060831282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,768,0.005526222288608551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,8192,0.010672888822025724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,4096,0.07391022311316596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,8192,0.052220443884531655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,32,0.008648888932334052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,1536,0.021016889148288306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,5120,0.22721955511305067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,8192,0.029839999145931665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,12288,0.03100088900989956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,1536,0.02238133384121789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,3072,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,7168,0.010971555279360877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,16384,0.5754888852437338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,768,0.005656888915432825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,2560,0.022257778379652236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,3584,0.008316444853941599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,256,0.005933333188295364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,6144,0.010423111418883005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,10240,0.020789333515697055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,7168,0.041971554358800255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,4096,0.010750222537252637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,128,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,6144,0.011761777930789523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,5120,0.01014666673209932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,1536,0.006226666685607698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,6144,0.05040622088644239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,3584,0.0472871098253462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,12288,0.037279999918407865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,65536,0.0759155551592509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,128,0.004880889008442561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,65536,0.4941102133856879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,1024,0.005283555636803309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,4096,0.016467556357383728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,10240,0.016881777180565726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,4096,0.014630221658282809
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,3584,0.007324444750944774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,16384,0.027912888262006972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,7168,0.10141511095894708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,1024,0.007843555675612556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,1024,0.0031973332580592898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,12288,0.021306667062971327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,2048,0.03225244416130914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,2048,0.023020444644822016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,6144,0.03194311261177063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,128,0.0028444443725877335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,1024,0.003909333298603694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,3072,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,12288,1536,0.015915556086434256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,3584,0.09020888805389404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,256,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,4096,0.010006222460005019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,64,0.009715555442704095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,512,0.021920889616012573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,6144,0.012994666894276937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,8192,0.018553778529167175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,128,0.012270222107569376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,65536,1.1295208401150174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,2560,0.007744000189834171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,8192,0.22365776697794595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,768,0.06916177935070462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,12288,0.011746666497654386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,16384,0.035762667655944824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,4096,0.11279377672407363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,1536,0.045066667927636035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,10240,0.02269866731431749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,12288,0.017179555363125272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,16384,0.03485511077774896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,128,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,10240,0.010682666467295753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,1024,0.006191111273235745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,7168,0.29843823115030926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,5120,0.0127368892232577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,8192,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,7168,0.03298222356372409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,6144,0.008000888758235509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,768,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,768,0.002826666666401757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,512,0.011961778004964193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,5120,0.019709333777427673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,6144,0.0192586663696501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,1536,0.0038488888078265717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,6144,0.0748222205373976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,65536,0.0860675573348999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,128,0.005544888890451855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,1536,0.006959111326270633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,7168,0.03604088889227973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,4096,0.022437334060668945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,256,0.006614222294754452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,6144,0.014145778285132514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,16384,0.02545333405335744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,65536,0.05986844168768989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,3584,0.015328889091809591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,3584,0.02015911042690277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,65536,0.08007111152013142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,5120,0.01612977849112617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,7168,0.025616889198621113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,3584,0.007296889192528195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,1536,0.01629155543115404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,12288,6144,0.04531822270817227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,16384,0.056447108586629234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,6144,0.004926222066084544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,4096,0.010804444551467896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,256,0.016520000166363187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,2048,0.006559111177921295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,2048,0.009389333426952362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,4096,0.012246222131782107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,1024,0.01852888862291972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,5120,0.013938667045699226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,3584,0.022672888305452134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,1024,0.013103110922707452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,256,0.005555555638339784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,768,0.0058711109062035876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,2048,0.020654223031467862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,12288,0.03607555561595493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,128,0.007258666886223688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,1536,0.010311110979980892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,768,0.005587555468082428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,2048,0.008649778034951951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,3072,0.012807110945383707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,10240,0.012113778127564324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,3584,0.056571556462181934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,12288,0.024847999215126038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,65536,0.28297334247165257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,6144,0.009448889229032729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,256,0.004923555586073133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,1024,0.01960711181163788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,2560,0.014477334088749357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,12288,0.05196711089875963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,3072,0.01032266683048672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,2048,0.013040888640615674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,256,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,6144,0.013036444783210754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,512,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,5120,0.009944000177913243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,4096,0.007337777978844113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,2048,0.016107555892732408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,3584,0.10375289122263591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,64,0.005409777578380372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,3072,0.004214222232500712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,10240,0.03818133473396301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,65536,0.05643910831875271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,4096,0.015418666932317944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,2048,0.009689778089523315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,128,0.0021199999998013177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,10240,0.012455999851226807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,5120,0.008656889200210571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,2048,0.010095111197895473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,6144,0.009685333404276106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,5120,0.009882666998439366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,1024,0.005543111099137201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,65536,0.08455466561847264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,512,0.009085333181752099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,7168,0.009655999640623728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,2048,0.031160887744691636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,5120,0.01755644381046295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,10240,0.027760000692473516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,2560,0.00795377790927887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,4096,0.01755466726091173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,768,0.016865778300497267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,512,0.09669599930445354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,8192,0.01070844464831882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,512,6144,0.005608888963858287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,3072,0.02421955598725213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,8192,0.028568890359666612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,5120,0.011843555503421359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,5120,0.016150222884284127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,256,0.002135111019015312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,16384,0.040892445378833346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,512,0.00328177772462368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,10240,0.08463111188676621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,5120,0.014408888088332282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,2048,0.00590844452381134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,3072,0.022008889251285132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,12288,0.2256284554799398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,768,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,7168,0.028396444188223943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,10240,0.01573155489232805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,65536,0.20018933879004586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,256,0.002558222247494592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,10240,0.0781715578503079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,7168,0.02251288957066006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,8192,0.04474311073621114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,256,0.026104888982243005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,1536,0.006592888798978593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,2048,0.013777777552604675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,32,0.0052151112920708126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,1536,0.0038151111867692736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,1536,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,2560,0.013316444224781461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,64,0.004906666775544484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,7168,0.026712889472643535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,3072,0.045704887972937695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,2560,0.03285600079430474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,2048,0.006646222124497096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,10240,0.01645777788427141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,256,0.014137778017255994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,5120,0.01442577772670322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,8192,0.01685599982738495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,32,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,12288,0.013057777451144325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,128,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,3072,0.011081777513027191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,1536,0.007802666889296637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,7168,0.15099999639723036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,256,0.006595555692911148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,128,0.004925333377387789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,16384,0.018744889232847426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,256,0.002467555511328909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,128,0.005534222142563925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,4096,0.0251857770813836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,7168,0.009930666950013902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1536,65536,0.03596355517705282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,8192,0.01404800017674764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,2560,0.006594666590293248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,512,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,768,0.005519999812046687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,16384,0.015178667174445258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,5120,0.008657777474986183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,3584,0.007680888805124495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,4096,0.008027555214034187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,5120,0.0325893329249488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,512,0.010327111515733931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,10240,0.018612444400787354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,3072,0.005391110976537068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,7168,0.24332356452941895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,2048,0.00666133314371109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,256,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,7168,0.014825777875052558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,6144,128,0.002529777793420686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,10240,0.18478222688039145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,512,0.01757688820362091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,3072,0.007213333414660559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,5120,0.011109333071443768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,10240,0.012335110869672565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,7168,0.01332800007528729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,512,0.0053280000057485365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,2048,0.0042568888101312844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,3072,0.008980444735950893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,5120,0.014977777997652689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,7168,0.034415112601386175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,6144,0.009000000026490953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,65536,0.0758133331934611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,64,0.0021928888228204516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,7168,0.02400622268517812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,1024,0.018556444181336295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,128,0.1974231137169732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,128,0.0053031109273433685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,7168,0.015127110812399121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,6144,0.18938399685753715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,16384,0.16324978404574925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,1536,0.004857777721352047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1536,10240,0.012447111308574677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,1536,0.018054222067197163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,7168,0.0143857780430052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,10240,0.04219111137919956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,8192,0.026920000712076824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,32,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,12288,7.911576165093316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,768,0.0038577777643998465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,3584,0.009134222235944536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,128,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,2560,0.011898666620254517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,16384,0.02480711042881012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,6144,0.02589599953757392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,1024,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,8192,0.02187555531660716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,6144,0.035311109489864774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,12288,0.017228444417317707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,3072,0.0107004443804423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,10240,0.05791111124886406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,1536,0.006245333287451003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,1536,0.45670398076375324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,32,0.0028648889727062653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,2560,0.010992888775136737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,1024,0.0035573335157500375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,5120,0.017668444249365065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,4096,0.01891288823551602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,5120,512,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,1024,0.02056533263789283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,768,0.009391110804345872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,12288,0.023021333747439917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,7168,0.011380444798204633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,1024,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,12288,0.020409777760505676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,6144,0.04273688793182373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,2560,0.012356444365448423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,512,0.04602311054865519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,6144,64,0.0052133335007561584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,16384,0.023374223046832617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,1024,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,2560,0.01239377756913503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,2048,0.0038497779104444715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,5120,0.028856890069113836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,1536,0.006228444476922353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,768,0.005903999838564131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,16384,0.02605155607064565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,128,0.013778666655222574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,1536,0.012730666332774691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,4096,0.008316444853941599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,256,0.007024889190991719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,1536,0.0058906666106647914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,16384,0.023331556055280898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,6144,0.009703999592198266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,64,0.0018586667461527719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,512,0.00490133340160052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,65536,0.11529510551028782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,64,0.0021582223061058256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,6144,0.008238222036096785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,1536,0.0038284444146686127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,2048,0.004182222402758068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,768,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,3584,0.00794577764140235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,3072,0.013743110828929476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,6144,0.015974221958054435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,2048,0.006577777779764599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,4096,0.00795199970404307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,128,0.007261333366235097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,1024,0.005984000033802456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,2048,0.006383111079533895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,6144,0.00904177791542477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,65536,0.025075554847717285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,6144,0.012670222255918713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,3072,0.02132355504565769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,16384,0.06949600246217516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,128,0.005062222066852781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,2560,0.08626577589246963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,3072,0.018914666440751817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,6144,0.0726613336139255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,128,0.0020151111400789684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,128,0.0021359999146726397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1536,2560,0.01259022206068039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,128,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,3072,0.013808888693650564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,4096,0.094595558113522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,256,0.026943999860021803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,3584,0.016760889026853774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,128,0.006573333508438534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,1024,0.0037715555065208008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,2560,0.007285333342022366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,12288,0.016564443707466125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,7168,0.013099555340078143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,12288,0.013793778088357715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,8192,0.011056889262464313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,128,0.03194488750563727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,1536,0.018209778600268893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,768,0.013119999733236102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,65536,0.25520621405707467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,3584,0.019007111589113872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,16384,0.022649778260125056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,3072,0.05899288919236925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,6144,0.029503110382292006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,16384,0.6057102415296767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,2560,0.0069191112286514705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,3072,0.007283555964628856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,6144,0.02313066687848833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,5120,0.06375110811657377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,5120,0.02587288949224684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,1536,0.014147554834683737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,512,0.0029066666546795103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,768,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,256,0.004920888692140579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,128,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,3584,0.00794400026400884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,2048,0.011374221907721626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,65536,0.07029422124226888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,5120,0.03788177834616767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,1536,0.005942222144868638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,1024,0.004591111093759537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,32,0.0069573331210348345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,512,0.006583999842405319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,1536,0.0243493335114585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,64,0.0027911110470692315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,5120,0.012831110921170978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,12288,0.04900000161594815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,256,0.014143110977278816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,12288,0.022463111413849726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,6144,0.13421777884165445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,128,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,1024,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,1536,0.006263111200597551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,512,0.0053084443012873335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,1536,0.004237333519591225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,2560,0.022406222091780767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,3584,0.010639111200968424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,768,0.015807110402319167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,1024,0.009015111459626092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,65536,0.06613155868318346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,12288,0.020815110868877836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,512,0.00851733320289188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,2048,0.00630044440428416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,1024,0.004230222354332606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,64,0.002191999927163124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,3072,0.03771911064783732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,2560,0.02847377790345086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,3072,0.007633777956167857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,12288,0.02602311140961117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,7168,0.017842666970358956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,5120,0.011743110915025076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,65536,3.299553765190972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,12288,0.02718044486310747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,12288,0.018704889549149405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,2560,0.00792711145348019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,2048,0.021811554829279583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,768,0.016536888149049547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,8192,0.15561155478159586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,768,0.00527022240890397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,256,0.01616533266173469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,8192,0.020831111404630873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,8192,0.037138667371537946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,5120,0.015311110350820752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,1024,0.007311111523045435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,64,0.006590222318967183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,256,0.004534222185611725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,2560,0.7365315755208334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,4096,0.009045333498054082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,10240,0.005212444398138259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,128,0.0918346643447876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,768,0.016167999969588388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,1536,0.00590133335855272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,128,0.009386666946940953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,2048,0.006565333240562015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,4096,0.01629688839117686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,16384,0.018545778261290658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,4096,0.011561777856614856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,10240,0.058746669027540416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,7168,0.08191111352708605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,3072,0.037364443143208824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,10240,0.0210817770825492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,7168,0.4703404638502333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,3072,0.05232888791296217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,12288,0.038795553975635104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,3584,0.006251555350091722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,128,0.00566133318675889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,10240,1.6835031509399414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,12288,0.09704355398813884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,5120,0.03903733359442817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,2560,0.01146933353609509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,12288,0.09259021944469875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,5120,16384,0.08027555545171101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,5120,0.01924888955222236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,256,0.007595555649863348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,2560,0.05298044284184774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,1536,0.02091377807988061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,1536,0.035409778356552124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,6144,0.010012444522645738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,2048,0.02767733401722378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,768,0.005948444621430502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,1024,0.011721777419249216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,1024,0.010737777584128909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,256,0.009594666461149851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,16384,128,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,4096,0.06405333015653822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,128,0.002833777831660377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,2560,0.044406221972571484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,16384,0.029318223396937054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,512,0.005610666755172942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,7168,0.0718088878525628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,16384,0.03073955575625102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,3584,0.014761777387724983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,1024,0.01389244364367591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,2560,0.006907555378145642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,7168,0.07842755317687988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,3584,0.011781333221329583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,2048,0.015478221906556023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,8192,0.02660977840423584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,12288,0.023073777556419373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,12288,0.014504889647165934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,2048,0.008681777450773451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,1536,0.01684711045689053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,4096,0.01684977776474423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,65536,1.625750223795573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,4096,0.10207111305660671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,16384,0.025033776958783466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,5120,0.03947822252909342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,10240,0.05347555544641283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,8192,0.011791110866599612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,128,0.010819555156760745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,4096,0.008368888662921058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,8192,0.06535644663704766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,768,0.014536889062987434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,3072,0.008580444587601556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,5120,0.008289777570300633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,2560,0.019179556104871962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,8192,0.02674755619631873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,8192,0.015792888071801927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,512,0.012033777932325998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,7168,0.014707555373509726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,3584,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,3584,0.03594222333696153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,2560,0.013399110900031196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,2560,0.008760000268618265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,512,0.018586667047606576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,3072,0.015610666738616096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,768,0.016908443636364408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,12288,0.32776000764634877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,2048,0.01870755520131853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,6144,0.0495671100086636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,32,0.0052764444715446895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,256,0.0025617778301239014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,7168,0.010059555371602377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,10240,0.005583111196756363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,64,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,512,0.006065777606434292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,6144,0.02092622220516205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,1024,0.006266666783226862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,10240,0.015789333317014907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,3072,0.027639999985694885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,5120,0.09451555543475681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,2560,0.009011555876996782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,3584,0.005942222144868638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,4096,0.010377777947319878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,8192,0.029260443316565618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,7168,0.037534223662482366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,65536,0.1555982165866428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,1024,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,1024,0.0192266669538286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,2560,0.010566222170988718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,768,0.00867022242810991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,256,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,128,0.006625777731339137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,16384,0.07457422547870211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,256,0.031494223409228854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,6144,0.06172177526685926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,1536,0.005949333310127258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,2560,0.02008800043000115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,8192,0.02088444431622823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,128,0.0053048887186580235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,2048,0.006238222122192383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,512,0.015156444576051501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,4096,0.06231644418504503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,10240,0.01835911141501533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,5120,0.03681866659058465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,1536,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,4096,0.010989333192507425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,1024,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,6144,0.019898666275872123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,10240,0.024001777172088623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,10240,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,768,0.005683555371231503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,6144,0.014132445057233175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,10240,0.0310204459561242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,128,0.005295111073387994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,3584,0.007318221860461765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,6144,0.029475556479560003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,3584,0.031268444326188825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,4096,0.020065777831607394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,16384,10240,0.05379111236996121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,8192,0.01957155598534478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,7168,0.01711466742886437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,12288,0.3616480032602946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,256,0.01651733285850949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,256,0.008197333249780867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,16384,0.10537066724565293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,12288,0.03673600157101949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,1024,0.0075911109646161394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,64,0.009355555805895064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,3072,0.041513777441448636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,1024,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,16384,1024,0.07834133174684313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,8192,0.010671999719407825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,16384,0.047332445780436196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,5120,5120,0.10228355725606282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,12288,0.041287110911475286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,2560,0.013131555583741931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,16384,0.03183999988767836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,65536,0.09827555550469293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,1536,0.006385777973466449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,256,0.0024702221983008915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,1536,0.0034942221310403612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,1024,0.018923555811246235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,1536,0.006270222365856171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,7168,0.004853333450025982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,64,0.04492888847986857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,16384,0.031111111243565876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,768,2560,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,768,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,16384,0.015789333317014907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,5120,0.012098666694429187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,65536,0.008676444490750631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,1536,0.005960000058015187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,64,0.0017884444031450485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,16384,0.09311644236246745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,12288,0.1042720013194614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,256,0.0018773333479960759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,512,0.013432888521088494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,1536,0.009658666948477427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,2560,0.02888622217708164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,2560,0.03327200147840712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,10240,0.013501333693663279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,16384,0.01720533271630605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,12288,0.058464891380733915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,2048,0.00453333349691497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,512,0.005544888890451855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,512,0.006337777607970768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,16384,0.031121777163611516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,12288,0.016300444801648457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,32,0.04515022370550367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,512,0.04036622246106466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,512,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,10240,0.014444443914625378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,3584,0.014200000299347771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,3072,0.00488711107108328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,4096,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,3584,0.007265778051482306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,10240,0.11097511317994858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,128,0.00962755580743154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,128,0.0028808888875775864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,65536,0.07528622282875909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,6144,0.01590577761332194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,1024,0.0031928889867332247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,7168,0.011380444798204633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,7168,0.011396444506115384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,10240,0.012100444071822695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,1536,0.016941333810488384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,4096,0.11070222324795193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,6144,0.013593778014183044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,3072,0.004371555729044808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,2048,0.007147555549939473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3584,128,0.004880889008442561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,7168,0.02678844498263465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,256,0.01585155559910668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,16384,0.006635555376609166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,5120,0.009110222260157267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,256,0.002185777864522404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,12288,0.08121422264311048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,4096,0.005974222388532426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,4096,0.0058453331391016645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,3584,0.006639110959238476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,8192,0.13349689377678767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,7168,0.010747555229398938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,1024,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,10240,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,1024,0.005967999911970562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,10240,0.21891821755303276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,768,0.005929777605666055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,16384,0.07466933462354872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,16384,0.029440889755884807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,7168,0.03625599874390496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,65536,0.1210835509830051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,1024,0.0057839999596277876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,768,0.008285333712895712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,7168,0.013112000293201871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,768,0.014505777094099255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,256,0.016805332567956712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,7168,0.013402666482660504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,1024,0.01723911033736335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,512,0.007948444121413762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,256,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,3072,0.012491555677519904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,65536,2.4684950510660806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,16384,0.047694222794638745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,2048,0.01923644377125634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,12288,0.0392506652408176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,3072,0.010046222143703038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,4096,0.008298666940795051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,512,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,3072,0.007315555380450354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,2560,0.00979377743270662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,65536,0.06797955433527629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,16384,0.015477332803938123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,3072,0.007259555160999298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,3072,0.013526221944226159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,12288,0.01832799944612715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,256,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,3584,0.03642311029964023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1024,768,0.015261333849694995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,4096,0.007726222276687622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,256,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,2048,0.014493332968817817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,5120,0.014175999495718213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,6144,0.013688888814714221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,5120,0.029806223180558946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,65536,0.06703911225001018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,8192,0.014468444718254937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,8192,0.011033777561452655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,128,0.014099554883109199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,6144,256,0.005249777601824866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,3072,0.030885332160525854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,16384,0.005877333382765452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,10240,0.052608887354532875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,12288,0.027830223242441814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,6144,0.03158222304450141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,2048,0.006274666637182236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,2048,0.08079466554853651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,512,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,7168,0.009659555223253038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,128,0.004869333157936732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,16384,0.020620443754725985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,256,0.010040889183680216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,3072,0.032623112201690674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,512,0.004399999976158142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,10240,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,768,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,3584,0.008019555773999956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,3072,0.01124622258875105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,6144,6144,0.03490399983194139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,7168,0.012158222496509552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,12288,0.07636355691485934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,3072,0.017151110702090792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,3584,0.08031821913189359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,256,0.002467555511328909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,3584,0.007626666790909237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,65536,0.0451546675629086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,2048,0.014767110347747803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,2560,0.01016533292002148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,8192,0.014753777119848462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,512,0.004938666605287128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,256,0.004910222358173794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,8192,0.03778310947948032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,8192,0.04200266798337301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,6144,0.18963466750250924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,3584,0.011332444846630096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,256,0.002529777793420686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,1536,0.005856888989607493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2560,128,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,3584,0.059471997949812144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,8192,0.03270488977432251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,4096,0.024655999408827886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,6144,0.009927111367384592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,768,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,65536,0.21519111262427437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,256,0.004951999833186467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,256,0.0021893332401911416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,8192,0.018547554810841877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,3584,0.018272888329293992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,7168,0.014719110396173267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,256,768,0.005633777628342311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,10240,0.11086489094628228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,5120,0.014429334137174817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,6144,0.07837600178188749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,3584,0.05110488997565376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,512,0.003237333355678452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,512,0.00656088896923595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3584,3072,0.007133333219422235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,10240,7168,0.032983111010657415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,4096,0.01200977795653873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,8192,0.030085331863827173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,6144,0.011759111450778114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,65536,0.0709164473745558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,256,0.01645866698688931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,3584,0.02942844563060337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,2560,0.006337777607970768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,5120,0.017881777551439073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,64,0.016905777984195285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,128,0.0020248889923095703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,1024,0.007659555309348636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,3584,0.007592000067234039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,1536,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,16384,0.940383063422309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,6144,0.019827556278970506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,4096,0.007941332956155142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,8192,0.027552889453040227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,4096,0.02001066671477424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,8192,0.017280000779363845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,128,0.004928888960017098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,256,0.007333333293596904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,10240,0.011804444094498953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,7168,0.020356444848908316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,2048,0.006611555400821898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,512,0.005681777579916849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,6144,0.022112889422310725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,2048,0.017588444881969027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,32,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,10240,0.04190311166975233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,512,0.002502222235004107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,8192,0.014309333430396186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,10240,4096,0.02492977844344245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,65536,0.25729510519239635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,1024,0.016879111528396606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,16384,0.022053332792388067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,7168,0.015455999308162265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,32,0.00796977761718962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,6144,0.009487999810112847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,8192,0.02162577708562215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,8192,0.012056888805495368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,1536,0.07066666417651706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,64,0.0018328889790508484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,4096,0.00793333351612091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,7168,0.004912888838185204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,8192,0.029885331789652508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,3072,0.004214222232500712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,7168,0.012168000141779581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,768,0.006979555719428592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,128,0.004934222333961063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,7168,0.018295110927687753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,1024,0.011749332977665795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,2048,0.006608888920810487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,6144,0.047740442885292896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,8192,0.021928000781271193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,128,0.002163555473089218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,12288,0.013283555706342062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,512,0.00798044436507755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,128,0.005221333354711533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,4096,0.007725333174069722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,768,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,768,0.010879110958841113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,32,0.0048942222363419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,8192,0.010012444522645738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,7168,0.08585066927803887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,512,0.0024951110697454875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,512,0.02026844521363576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,1024,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,8192,0.03864177730348375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,512,0.002895111011134254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,32,0.007335999773608313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,10240,0.056497779157426625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,65536,0.3117431004842122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,16384,16384,0.08868710862265693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,6144,0.035972442891862653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,5120,0.09652088748084174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,4096,0.04702666733000013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,12288,0.007091555330488417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,768,0.00675999994079272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,768,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,6144,0.008984888593355814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,7168,0.12942666477627224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,1024,0.016154666741689045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,12288,0.02978666623433431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,2560,0.05203466614087423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,12288,0.005121777868933148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,12288,0.13166488541497126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,7168,0.014645333091417948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,6144,0.1355395581987169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,768,0.010928000013033548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,10240,0.017198221551047433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,5120,0.015172445111804538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,128,0.005081777771313985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,7168,0.015260444747077094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,2048,0.040349334478378296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,1536,0.004268444246715969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,1536,0.01684177749686771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,12288,0.01713244451416863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,65536,0.13022132714589438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,256,0.004884444591071871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,2048,0.011350221931934357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,2048,0.04172799984614054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,2560,0.028154667880800035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,12288,256,0.006020444548792309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,7168,0.08373155858781602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,8192,0.04800533254941305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,65536,0.029534220695495605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,8192,0.028712001111772325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,256,0.0024862223201327855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,65536,0.23544443978203666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,1024,0.005562666803598404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,4096,0.007653333246707916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,768,0.005601777798599667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,7168,7168,0.020899555749363367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,8192,0.03412444392840067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,2048,0.0069102222720781965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,1536,0.006098666538794835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,3584,0.017678222722477384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,128,0.018350222044520907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,2048,0.01998044384850396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1536,12288,0.05063377817471822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,3584,0.031509333186679415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,6144,0.009453333086437648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,16384,0.7865937550862631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,2048,0.01629599928855896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1536,8192,0.02583377725548214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,3584,0.00794311116139094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,12288,0.0185217774576611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,4096,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,3072,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,2048,0.014569777581426831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,32,0.0018542222678661346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,65536,0.12488267156812881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,2048,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,3584,0.012336889074908363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,7168,0.07894755734337701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,768,0.01586844523747762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,2560,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,3072,0.00998488896422916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,3584,0.011597333682907952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,128,6144,0.015843555331230164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,16384,0.017208000024159748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,7168,0.031665778822369046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,1536,0.007393777370452881
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,5120,0.014103111293580798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,16384,0.03951199849446615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1536,4096,0.01015644437736935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,12288,0.027181333965725366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,6144,0.004553777890072929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,7168,0.6015360090467665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,512,0.013697778185208639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,4096,0.02453244394726223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,16384,0.0810355544090271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,2560,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,5120,0.008635555704434713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,10240,0.037244445747799344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,65536,0.08255200253592597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,12288,0.02290311124589708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,256,0.01443911095460256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,4096,0.009317333499590555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,2048,0.015116444892353482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,6144,0.1630506647957696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,6144,0.009296000003814697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,3072,0.02641777859793769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,1536,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,5120,0.02095111045572493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,7168,0.01588444411754608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,512,0.006298666612969504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,2560,0.007607999775144789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,3584,0.007526222202512953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,8192,2560,0.009389333426952362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,10240,0.027483556005689833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,128,0.011408000356621213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,5120,0.021135111649831135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,7168,0.010991999672518836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,64,0.007991111112965478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,12288,0.04133866561783685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,2048,0.021319111188252766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,2560,0.01574222246805827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,16384,0.04062133365207248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,512,0.012740444805887012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,3584,0.008284444610277811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,5120,0.016844444804721408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,128,0.004951111144489712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,7168,0.023338667220539514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,128,0.005210666606823604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,8192,0.01752800080511305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,5120,0.024158222807778254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,4096,0.024039111203617517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,5120,0.0573671129014757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,5120,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,8192,0.017492444978819955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,8192,0.7741235627068414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,4096,0.15578399764166936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,1024,0.0031822222388452957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,3072,0.008032889001899296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,6144,0.1620595587624444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,512,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,1536,0.006621333460013072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,1024,0.003173333282272021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,5120,0.010035555395815108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,12288,0.017533333765135873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,768,0.05933688746558296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,256,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3584,512,0.013776889277829064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,12288,0.07204444540871514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,128,0.002505777817633417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,8192,0.08726666371027629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,16384,0.024668445189793903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,6144,0.05264800124698215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,10240,0.1095199982325236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,7168,0.018591110905011494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,8192,0.03542666633923849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,12288,0.785214212205675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,128,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,1024,0.015464888678656684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,128,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,10240,0.022188444932301838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,128,0.002136888810329967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,2048,0.00832711077398724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,10240,0.02741066614786784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,512,0.020250666472646926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,128,0.00555288874440723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,65536,0.06004533502790663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,1536,0.012474666866991254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,256,0.014837332897716098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,3584,0.041843556695514254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,3072,0.007295110987292395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,1536,0.011393778026103973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,128,0.002128888956374592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,65536,0.1071102221806844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,16384,1024,0.025470222036043804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,512,0.002480888946188821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,64,0.005314666777849197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,2048,0.1591093275282118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,768,0.02537155482504103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,7168,0.012847999731699625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,5120,0.008979555633332994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,3584,0.006634666687912411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,3072,0.025639111797014873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,1024,0.013047110703256396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,32,0.006585777633719974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,12288,0.06551200151443481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,10240,0.01516177753607432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,3072,0.01707466608948178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3072,256,0.015801777442296345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,32,0.002830222249031067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,16384,0.023207111491097346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,6144,0.06320355335871379
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,256,0.012724444270133972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,12288,0.010302222437328762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,2560,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,12288,0.015841777125994366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,128,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,10240,0.01640888883007897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,65536,0.029044446018007066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,512,0.007345777418878343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,1024,0.016184000505341422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1024,7168,0.07166488965352376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,512,0.00425155543618732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,1536,0.011390222443474663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,6144,0.01220355596807268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,768,0.020608888732062448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,12288,0.01718755563100179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,2560,0.005236444373925527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3072,128,0.014126222994592456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,1024,0.005850666513045629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,12288,0.013439111411571503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,3584,0.02032533288002014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,768,0.005609777652555042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,3072,0.007027555671003129
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,3072,0.007652444144090016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,7168,0.006266666783226862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,2048,0.013080000049538083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,16384,0.019966221517986722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,4096,0.0105262224872907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,512,0.015186667442321777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,2560,0.008976889153321585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,2048,0.028147555059856836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,8192,0.01753155555990007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,2048,0.00573955559068256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,512,0.0069546666410234236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,16384,0.0199306673473782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,256,0.012357333468066322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1536,1024,0.003237333355678452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,5120,0.00849422232972251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,2560,0.022038223014937505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,256,0.004863111095296012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,6144,0.016403555870056152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,1024,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,4096,0.007941332956155142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,64,0.009304888546466827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,10240,0.013415999710559845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,3072,0.006918222126033571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,128,0.0021991110924217436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,1024,0.004991111242108875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,3072,0.010964444114102257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,128,0.006229333165619109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,1024,0.0055440002017551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,3072,0.0073315559162033936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,6144,0.049638221661249794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,3072,0.0052435555391841466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,3584,0.00703022215101454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,7168,0.09163466427061294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,5120,0.021926222576035395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,7168,0.010377777947319878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,768,0.007319110963079665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,768,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,128,0.0029066666546795103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,64,0.0026133333643277488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,5120,0.02647200061215295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,128,0.0021742222209771476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,512,0.013781333135233985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,5120,0.059021333853403725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,256,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,128,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,12288,0.017183999220530193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,3072,0.0147706667582194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,2560,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,2560,0.007050666544172499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,65536,0.03074044320318434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,4096,0.12197332912021214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,32,12288,0.0069617778062820435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,3584,0.03231822119818793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,65536,0.2511857880486382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,8192,0.011819555527634092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,2560,0.04105866617626614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,3072,0.00828888929552502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,512,0.015484443969196744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,768,0.00555644432703654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,7168,256,0.004941333499219683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,16384,0.015848888291252982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,8192,0.012853333519564735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,3072,0.004616888860861461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,5120,0.029856890439987183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,5120,0.060368888907962374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,128,0.01035288886891471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,768,10240,0.01330311099688212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,2048,0.29927200741238064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,12288,5120,0.17321511109670004
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,64,0.008295999632941352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,512,0.0052933332820733385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,16384,0.04577599962552389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,2048,0.03592355714903937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,8192,0.011006222003036074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,65536,32,0.007247111035717859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,1536,0.029478222131729126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,1536,0.004350222233268949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,128,0.004920000003443824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,256,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,768,5120,0.021031111478805542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,2048,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,5120,0.8241377936469184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,7168,0.026339555780092876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,32,0.002120888895458645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,10240,0.018401778406567044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,768,0.006584888945023219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,2048,0.00867022242810991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,12288,0.027452444036801655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,256,0.002458666761716207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,1536,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,1536,0.003924444317817688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,3072,0.07630222373538546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,32,0.024079999989933435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,65536,0.11693510744306777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,256,256,0.004898666507667965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,512,0.013824889229403602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,8192,0.03372800019052293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,2048,0.01756977703836229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,768,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,1024,0.0031600000543726813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,2048,0.005320000151793162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,768,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,5120,0.026439110438028973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,5120,0.021288888321982488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,3584,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,1024,0.019353777170181274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,12288,0.03671377897262573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,5120,0.008623110751310984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,1536,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,12288,0.03397511111365424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,64,0.004257777912749184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,1024,0.009737778041097853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,512,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,3584,0.007320000065697565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,8192,0.051036443975236684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,65536,16384,0.28363111284044057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,65536,0.17804977628919813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,7168,0.27156978183322483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,3584,0.016020443704393175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,1024,0.02163200080394745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,5120,0.034260445170932345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,4096,0.027842667367723253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,2560,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,8192,0.05755111243989733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,6144,1536,0.014113777213626437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,4096,0.012082666986518435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,3072,0.030623998906877305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,3584,0.007248889240953658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,6144,1536,0.006224888894293044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,512,0.002477333363559511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,128,0.002124444478087955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,12288,0.044481777482562594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,256,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,65536,0.007177777588367462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,4096,0.00832711077398724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,768,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,1024,0.019882667395803664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,128,0.003510222252872255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,10240,0.009533333281675974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,128,0.016744888491100736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,2048,0.007300444775157505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,65536,0.24157510863410103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,512,0.005207111024194294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,128,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,12288,0.2598035600450304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,65536,0.05790222353405423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,12288,0.013752000199423896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,3584,0.03499644332461887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,2560,0.045583999819225736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,12288,0.013247110777431063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,4096,0.01994311147265964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,2560,1.5664400524563258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,10240,0.03706577751371596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,6144,0.08585688802931045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,512,0.015494222442309061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,64,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,2048,0.03248000144958496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,7168,128,0.00868266655339135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,1536,0.011848889291286469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,7168,0.05558311276965671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,2560,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,3072,0.0317813323603736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,10240,0.01274311128589842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1024,3584,0.0075866662793689305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,32,0.00793333351612091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,1024,0.07897510793473986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,256,0.0058853332367208265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,2048,0.02699111070897844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,1024,0.005522666705979242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,5120,0.020946666598320007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,128,0.002489777695801523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,3072,0.01034044474363327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,128,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,12288,0.06452533271577623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,4096,0.007927999728255803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,1536,0.006595555692911148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,7168,0.023123555713229712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,6144,0.012816888590653738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,1536,0.00591466658645206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,1536,0.02749066717094845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,3072,0.013096888860066732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,3584,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,5120,0.008330666356616551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,5120,0.011843555503421359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,4096,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,2048,0.009280888570679558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,1024,0.015463110473420886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,3584,0.007322666545708974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,768,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,3072,0.009286222358544668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,65536,0.10432266526752049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,1536,0.006236444330877728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,1024,0.016532444291644625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,2560,0.006626666833957036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,2560,0.00658044425977601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,768,0.39450666639539933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,10240,0.030441777573691473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,1536,0.0467564430501726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,512,0.005969777703285217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,16384,0.058281779289245605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,65536,0.0980186661084493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,65536,0.06325688627031115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,256,0.003882666842805015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,1536,0.01684799955950843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,7168,0.026457778281635706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,1536,0.015525332755512662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,3584,0.011408888631396823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,8192,0.033062222931120135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,4096,0.03432888785998026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,4096,0.007965332931942409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,12288,0.024199111594094172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,8192,0.014360888136757744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,2048,0.006551110910044775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,4096,0.029057777590221826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,65536,0.1435111098819309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,64,0.0022159999029503926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,10240,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,256,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,5120,0.011937778029176923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,32,0.005607999861240387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,64,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,768,0.005187555733654234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,128,0.007272000114123027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,10240,0.018222222725550335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,8192,0.029707554313871596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,128,0.011744889120260874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,4096,0.09947022464540269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,16384,0.1682551172044542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,3072,0.007314666277832455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,256,64,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,2048,0.0064284445510970215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,16384,0.01868711080816057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,768,0.030193779203626845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,3072,0.02033333314789666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,2560,0.006650666809744305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,2048,0.004209777961174647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,64,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,3584,0.04626755581961738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,2048,0.024717332588301763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,16384,0.01585155559910668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,4096,0.04172177778349983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,512,0.01993955506218804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,2560,0.14941244655185276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,6144,0.06805600060356988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,2048,0.006647999915811751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,1536,0.0173999998304579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,1536,0.016184000505341422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,16384,0.017997332745128207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,256,0.002187555655837059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,1024,0.0031439999325407874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,4096,0.06572800212436251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,3584,0.006617777877383762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,10240,0.018948444061809115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,5120,0.009400889277458191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,2560,0.03258400162061056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,1536,0.009910222556855943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,128,0.0022195554855797025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,1024,0.00693866651919153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,512,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,1024,0.017267554998397827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,3584,0.01666755477587382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,1536,0.007601777712504069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,6144,0.012723555167516073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,8192,0.3182062307993571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,6144,0.01684266659948561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,3072,0.004221333397759331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,65536,0.1114693350262112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,128,0.010384000009960597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,64,0.0031973332580592898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,2560,0.0244186669588089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,12288,0.018539556198649935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,256,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,1536,0.010319111247857412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,8192,0.006583999842405319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,512,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,3584,0.00831288927131229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,7168,0.039705776506000094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,1536,0.006069333189063602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,3584,0.007262222468852997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,2560,0.014431110686726041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,32,0.005541333307822545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,10240,0.016758221719000075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,2048,0.06411377588907878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,2048,0.00832622249921163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,32,0.0052720002002186244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,5120,0.03127377894189622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,5120,0.018210666047202218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,1024,0.02234577801492479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,65536,0.14084711339738634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,6144,0.026986666851573523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,128,0.005899555567238066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2048,10240,0.025997334056430395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,8192,0.040785776244269475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,3584,0.0041653331783082755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,16384,0.028059555424584284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,5120,0.017418666018380057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,768,0.005417777846256892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,3072,0.037104888094796076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,768,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,256,0.014085332552591959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,128,0.0032017777363459268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,65536,0.07631288634406196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,12288,64,0.01698755555682712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,1536,0.01983022193113963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,1024,0.0057795556883017225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,768,0.0052284445199701525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,4096,0.01996977792845832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,2048,0.006279110908508301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3584,8192,0.02217777735657162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,512,0.006745777610275481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,2560,0.028825776444541082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,12288,0.02945244312286377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,4096,0.03123377760251363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,5120,0.012084444363911947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,3072,0.00461777796347936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,16384,0.6478382216559516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,12288,0.013989332649442883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,3584,0.02638133366902669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,128,0.012476444244384766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,512,0.003124444435040156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,7168,0.02474400069978502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,12288,0.015964445140626695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,128,0.004869333157936732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,12288,0.026357332865397137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,12288,0.020994666549894545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,7168,0.013796444568369122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,256,0.0021884443445338146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,1536,0.015819556183285184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,1024,0.03126577867401971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,128,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,512,0.003899555653333664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,7168,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,2560,0.0042444442709287005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,512,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,65536,0.08135200209087796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,64,0.016191111670600045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,2560,0.004869333157936732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,12288,0.13073599338531494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,16384,0.04301688737339444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,3072,0.007328000333574083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,65536,1.0948817994859483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,16384,0.022658665974934895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,6144,0.11498933368259007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,12288,0.02092355489730835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,4096,0.011434666812419891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,2048,0.00692622239391009
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,16384,0.0566577778922187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,2048,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,4096,0.014111111561457316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,12288,0.07983466651704577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,4096,0.006583111153708563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,2560,0.007641777396202087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,7168,0.022329777479171753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,2560,0.008287111090289222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,512,0.021331555313534204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,1024,0.003859555555714501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,128,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,2048,0.007246221933099959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,32,0.004547555413511065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,65536,1.0655253728230794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,128,0.00226488895714283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,2048,0.020057777563730877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,512,0.015091554986106025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,65536,1.4871128929985895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,128,0.0025351111673646504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,6144,0.005692444327804778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,1024,0.05307911170853508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,768,0.0038622220357259116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,1536,0.0052151112920708126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2048,2048,0.023511999183230933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1024,8192,0.06199644671546089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,128,0.05510666635301378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,12288,1024,0.009947555760542551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,1536,0.008271999657154083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,1536,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,1024,0.01512888901763492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,32,0.007595555649863348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,65536,0.1307351059383816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,512,0.007983999947706858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,512,0.012396444049146442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,1024,0.005567111074924469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,5120,0.012595555848545499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,256,0.014814222852389017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,7168,0.01552888916598426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,12288,0.20806577470567492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,768,0.00564444437623024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,4096,0.008013333711359236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,256,3584,0.005600000007285013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,12288,0.05964622232649061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,256,0.017887110511461895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,512,0.006641777853171031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,256,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,3072,0.023580445183648005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,1536,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,256,0.013801777528391944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,5120,0.00831288927131229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,3584,0.0041777777175108595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,5120,0.012736000120639801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,1024,0.0042133331298828125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,16384,0.3644373416900635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,1024,0.015861334072219003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,1536,0.012806221842765808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,3072,0.007000000112586551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,2048,0.006656888872385025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,3584,0.08333244588640001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,32,0.014056889547242058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,768,0.014677332507239448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,32,0.0018533333722088074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,16384,0.06755999724070232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,12288,0.2656888961791992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,2560,0.004917333523432414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,3584,0.021024000313546922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,12288,0.0322728885544671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,8192,0.00553688903649648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,10240,0.052417778306537204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,3072,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,512,0.0032266666077905228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,4096,0.007650666766696506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,8192,0.08615022235446507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,65536,0.3956577777862549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,16384,0.02810488806830512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,512,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,10240,0.20655288961198595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,768,0.13954044712914362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,6144,0.14358577463361952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,5120,0.012247111234400006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,5120,0.11641244093577068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,768,0.011254222028785281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,6144,0.012230222423871359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,128,0.036737776464886136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,768,0.003516444315512975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,512,0.00657155571712388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,10240,0.017839999662505258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,5120,0.013863999810483722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,3584,0.009691555466916826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,768,0.005531555662552516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,7168,0.02188266648186578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,16384,0.023357333408461675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,768,0.02555822167131636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,768,0.024505777491463557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,256,0.0018737777653667664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,768,64,0.0018577778504954444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,2048,0.01072266697883606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,7168,1536,0.006867555694447623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,5120,0.024068444967269897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,4096,0.005320000151793162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,5120,0.009743111001120673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,5120,0.03810755411783854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,768,0.01662133302953508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,1024,0.014340443743599786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,12288,1.8354835510253906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,4096,6144,0.01277955538696713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,2048,0.01200711147652732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,6144,0.013208000196350945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,128,0.01719288859102461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,6144,0.009353777600659264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,7168,0.04702488912476433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,2560,0.01640888883007897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,10240,0.023783110909991797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,3584,0.008996444443861643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,12288,4096,0.011773332953453064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,512,0.016501333978441026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,3072,0.006976000136799282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,16384,0.07276178068584867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,4096,1536,0.006182222316662471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,65536,0.04411022199524773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,16384,0.08006044228871663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,10240,0.10145066844092475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,8192,0.013640888863139682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,512,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,512,0.00592533333433999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,6144,0.024685333172480266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,5120,0.0042568888101312844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,2560,0.008976000050703684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,2560,0.006904888898134232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,512,0.01518311103185018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,16384,0.015094222293959724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,3072,0.03801511062516107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,2560,0.02084000077512529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,1536,0.007975111405054728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,16384,0.05752444267272949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,2560,0.005193777796294955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,5120,5120,0.030211554633246526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,128,0.005241777747869492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,6144,0.08810666534635757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,1536,0.029108444849650066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,65536,0.05413155423270332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,3072,0.01584533353646596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,32,0.0018151110659042995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,32,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,16384,0.09628533654742771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,2048,0.021545777718226116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,1024,0.007294221884674496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,4096,0.011385777758227454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,3584,0.01074577785200543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,6144,0.016418667303191293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,1536,0.007925333248244392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,7168,0.04235555397139656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,3584,0.03712800145149231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,16384,0.03414400087462531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,12288,0.020280889338917203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,16384,0.026230222649044458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,16384,0.031888888941870794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,2560,0.007151111132568783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,64,0.00360355567600992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,32,0.005946666830115848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,65536,0.11450489362080891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,2560,0.013447999954223633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,5120,0.010040000081062317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,8192,0.005595555735958948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,12288,0.023744000328911677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,2048,0.006640000061856375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,3584,0.011419555379284753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,1024,0.0306346681382921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,16384,2560,0.01684711045689053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,128,0.013453333742088742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,16384,0.07245510816574097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,1024,0.009361777868535783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,16384,0.036976890431510076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,5120,0.026410667432679072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,128,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1024,512,0.008671999805503422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,768,0.0110880004035102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,8192,1536,0.023660444551044043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,2048,0.004537777768241035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,7168,0.037905779149797224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,12288,0.02438844409253862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,5120,0.011374221907721626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,16384,0.015453333655993143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,256,0.01514755520555708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,128,0.004915555732117759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,8192,0.015479111009173922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,1536,0.0041866666740841335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,65536,0.5236346456739638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,5120,0.008692444198661381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,2560,0.068996442688836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,2560,0.013546666337384118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,6144,0.019916445016860962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,2048,0.015473778049151102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,3072,0.007310222420427535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,16384,0.028138667345046997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,2048,0.04558311237229241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,256,0.004908444566859139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,1024,0.029503110382292006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,3072,0.006952889263629913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,5120,0.022072000635994807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,512,0.015119110544522604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,10240,0.01033333357837465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,1024,0.006236444330877728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,128,0.013766222529941134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,7168,4096,0.011472889118724398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,4096,0.007770666645632849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,1024,0.002983111060327954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,65536,0.1021804412206014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,128,0.00527022240890397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,3072,0.037010666396882795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,128,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,256,0.0031928889867332247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,8192,2048,0.008316444853941599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,512,0.019333332777023315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,65536,0.12678933143615723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,1536,0.12324178218841553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,1024,0.022486221459176805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,3584,0.11895999643537734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,16384,0.025383111503389146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,3072,0.007243555453088548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,128,0.00508355556262864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,16384,0.3571466604868571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,2048,0.477493339114719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,256,0.011719110939237805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,3072,0.014097778333557976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,12288,0.03624799847602844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,4096,0.029401779174804688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,256,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,16384,0.025018667181332905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,512,0.0024826667375034755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,12288,0.012109333442317115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,32,0.0024844445288181305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,3072,0.016745777593718637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,5120,0.021201777789327834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,32,0.0021600000974204806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,64,0.0018542222678661346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,10240,0.08277689086066352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,2048,0.01921955578856998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,6144,0.55640533235338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,12288,0.02805066770977444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,65536,0.18310044871436226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,8192,0.041931556330786816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,4096,0.028435554769304063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,5120,0.009712888962692684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,6144,2560,0.0069306666652361555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,256,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,1536,0.00793511089351442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,768,0.005537777725193236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,2560,0.016145777371194627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,3072,0.0069368887278768755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,7168,0.09848266839981079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,16384,0.02437422176202138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,7168,0.024127110838890076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,6144,0.020592888196309406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,3072,0.022288888692855835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,3584,0.010951999988820819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,16384,0.31531643867492676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,2048,0.1497404442893134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,512,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,1536,0.006647111227114995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,64,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,768,0.015456888410780164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,10240,0.022329777479171753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,64,0.005251555393139522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,5120,0.008961777720186446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,6144,0.03209599852561951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,16384,0.025507556067572698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,16384,0.017509332961506315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,3072,0.007674666742483775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,3072,0.02403733299838172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,256,0.014736000034544202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,256,0.015110222829712762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,3072,0.022999111149046157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,512,0.00300799993177255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,2048,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2048,3072,0.0058711109062035876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,4096,0.014783999986118741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,3072,0.020617778102556866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,3072,0.03795555565092299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,7168,0.012050666742854647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,65536,0.08594755331675212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,3072,0.06491644514931573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,128,0.002157333410448498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,12288,0.020083554916911654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,1536,7168,0.011693333586057028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,256,0.007990222010347579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,8192,0.32036622365315753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,7168,16384,0.7898080084058973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,3072,0.02572533322705163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,768,0.004214222232500712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,16384,0.11566221714019775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,12288,0.9889893001980252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,3584,0.008296888735559251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,3584,0.014487110906177096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,2048,0.018271999226676095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,7168,0.04121155540148417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,12288,0.022613333331214056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,10240,0.03480622172355652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,10240,0.006291555447710886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,12288,0.011744889120260874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,512,0.01183111137813992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,10240,0.10582133134206136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,3584,0.028880000114440918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,1536,0.01514488955338796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,4096,0.032912890116373696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,3072,0.014444443914625378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,8192,0.012094222009181976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,3072,0.006976889239417181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,65536,0.07598844501707289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,768,0.04895200000868904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,7168,0.012739555703269111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,2560,0.019474666979577806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,7168,0.025535111625989277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,768,1024,0.013455111119482251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,2048,0.02605688903066847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,16384,0.015781333049138386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,512,0.002577777744995223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,768,0.018233777748213876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,8192,0.010909332997269101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,2048,0.004234666625658671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,6144,8192,0.16626400417751735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,128,0.01348088930050532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,2048,0.006291555447710886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,7168,0.007385777930418651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,16384,0.3020462195078532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,65536,0.05422666668891907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,4096,0.04621155394448174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,10240,12288,0.03321422139803568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,65536,7168,0.1262764400906033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3072,12288,0.02602311140961117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,512,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,128,0.005225777626037598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,10240,0.030216889248953924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,65536,0.1979937818315294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,5120,0.009017777939637503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,4096,0.017607111069891188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,6144,8192,0.025083555115593806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,7168,0.010053333308961656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,128,0.004873777843183941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,768,0.029691557089487713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,128,0.0140657772620519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,7168,0.00906311141120063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,16384,0.0785964462492201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,5120,0.018564444449212816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,10240,0.06359110938178168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,1536,0.05866489145490858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,5120,0.012357333468066322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,64,0.009700444837411245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,65536,0.08775999810960557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,65536,0.16597421964009604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,64,0.01105955574247572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,256,0.0023431111541059283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,4096,0.01625333395269182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,8192,0.19617511166466606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,5120,0.02834577692879571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,256,0.005600888695981767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,2560,0.006626666833957036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,3584,0.007282666862010956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,4096,0.009398221969604492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,512,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,768,0.01720177796151903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,512,0.005208888815508948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,128,65536,0.013526221944226159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,3072,0.007196444604131911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,2560,0.06402489211824205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,768,0.041532443629370794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,768,0.006280888699822956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,768,0.0031431110368834603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,16384,0.015462223026487561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,12288,0.013791110780504016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,8192,0.015975111060672335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,16384,0.02723466687732273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,5120,12288,0.19655911127726236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,10240,3584,0.014124444789356656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,65536,0.008338666624493068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,12288,0.0539173318280114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,256,0.00249422217408816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,768,0.0028248888750871024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,1536,0.003510222252872255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,7168,0.1280346711476644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,3584,0.02069244450993008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,256,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,3584,0.004598222259018156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,1536,0.015606222881211175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,65536,0.060989333523644336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,7168,0.047056890196270414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,3584,0.008608889248636033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,32,0.0018044445249769424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,32,0.007253333098358578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,7168,0.028951111767027114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,1024,0.014118222726715935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,2560,0.022375999225510493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,5120,0.008960888617568547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,7168,0.14779910776350233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,10240,0.14713066154056126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,4096,0.017895999881956313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,3072,0.023591111103693645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,6144,8192,0.014821334017647637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,10240,0.01683555543422699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,3072,0.019221333993805777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,3072,0.007425777614116669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,32,0.002161777681774563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,7168,0.01787644459141625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,7168,0.01646311084429423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,128,0.00489688871635331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,3072,0.005408888889683618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,5120,0.01035644445154402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,1536,3072,0.022464000516467627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,10240,16384,0.06347199943330553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,512,0.023249778482649062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,5120,0.028548445966508653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,4096,0.01773066653145684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,16384,0.1430222193400065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,4096,0.015835555063353646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,1024,0.035848889085981585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,2048,0.027599111199378967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,256,0.035649776458740234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,65536,0.06266933017306857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,2560,0.0064186664919058485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,32,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,3584,0.00657066661450598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,512,0.004856889032655292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,3072,0.0041999999019834726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,6144,0.013795555465751223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,1024,0.003612444632583194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,4096,0.008993777963850234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,1536,0.05274666680230034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,256,0.014127999544143677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,2048,0.006279110908508301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,3584,0.00795733349190818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,12288,0.06814044713973999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,1536,0.007982222570313348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,64,0.0028293333533737394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,32,0.04139644569820828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,8192,16384,0.04877066612243652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,4096,0.00794577764140235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,2560,0.040419555372662015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,6144,0.024548444483015273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,12288,0.15282932917277017
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,4096,0.033991999096340604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,1536,0.007596444752481248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,768,0.006256888724035687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,8192,0.07824266619152494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,12288,0.025221332907676697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,512,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,10240,0.020252444677882724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,256,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,4096,0.1465537812974718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,1024,0.004910222358173794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,1536,0.006134222365087933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,16384,0.023400889502631292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,5120,0.01387466655837165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,3072,0.09399822023179795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,4096,768,0.015581332974963717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,2048,0.0038480001191298165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,5120,0.00832177781396442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,2048,0.006341333190600078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,8192,0.009091555244392818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,768,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,512,0.011054221954610614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,65536,0.10990489191479152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,1024,0.005944889038801193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,6144,0.01205599970287747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,256,0.0032008888406885993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,8192,0.01037688884470198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,128,0.012345777617560493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,128,0.0022026666750510535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,3072,0.010133333504199982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,512,0.0032177778581778207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,8192,0.02126577827665541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,1024,0.004581333448489507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2560,65536,1.3059102164374459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,2560,0.04495822058783638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2048,512,0.013735111388895245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,3584,0.009700444837411245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,1536,0.014821334017647637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,6144,0.010039110978444418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,16384,0.02919911013709174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,10240,0.012103111379676394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,128,0.005882666756709416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,8192,0.0107342220014996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,6144,0.00976533359951443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,64,12288,0.005078222188684675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,3072,0.021762667430771723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,3584,0.020259555843141343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,7168,0.023358222511079576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,5120,0.01520088811715444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,8192,0.1884533299340142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,7168,0.013178666432698568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,2560,0.03583733240763346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3584,65536,0.20864444308810762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,3072,0.011398221883508893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,7168,0.04953866534762912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,7168,0.03773422373665704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,7168,0.010002666877375709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,1536,0.006784889019197888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,5120,0.008588444855478074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,16384,0.15630311436123318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,3072,0.007363555332024892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,6144,0.016879111528396606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,6144,0.032904889848497175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,12288,0.012736000120639801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,6144,0.14496800634596083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,7168,0.04636711213323805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,2048,0.006739555547634761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,65536,0.06505599949094984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,3072,0.27328088548448354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,768,0.002951111023624738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,6144,0.010433778166770935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,16384,0.32505067189534503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,2560,0.00592533333433999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,256,0.005966222120655908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,128,0.002509333400262727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,2560,0.01420177850458357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,8192,0.01107911103301578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,3072,0.007316444483068254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,8192,0.020238222347365487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,256,0.013991110854678683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,512,0.002567111204067866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,8192,0.01757333344883389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,16384,0.026329777306980554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,65536,128,0.01869244376818339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,6144,0.020634666085243225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,512,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,6144,0.18677866458892822
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,5120,0.012802666260136498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,10240,0.013951111170980664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,1024,0.005972444597217772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,2048,0.012855110896958245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,12288,0.016511999898486666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,512,0.00701333334048589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,65536,1.0911493301391602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,1024,0.003567111161020067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,16384,0.23496267530653211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,6144,0.015491555134455362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,10240,0.025937777426507738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,12288,0.10140711069107056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2048,4096,0.007622222105662028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,1024,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,65536,768,0.02420622275935279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,4096,0.02198933396074507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,64,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,256,0.05050399899482727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,65536,0.06389600038528442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,10240,5120,0.2931022114223904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,2560,0.004925333377387789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,512,0.0025253333151340485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,16384,0.04652355445755852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,512,0.01443199978934394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,64,0.010996444357766045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,768,5120,0.01722133325205909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,3584,0.019927110936906602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,1536,0.007283555964628856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,16384,0.026152888933817547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,7168,0.01352355546421475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,8192,0.05942666530609131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,3072,0.007599111232492659
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,16384,0.0333555539449056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,64,0.0021386666016446217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,6144,0.020430222153663635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,5120,0.008998221821255153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,4096,0.005229333208666907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,10240,0.03598577777544657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,512,7168,0.009679999616410997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,3072,0.007683555285135905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,7168,0.3143528832329644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,512,0.007486222518814935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,5120,0.018177777528762817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,8192,0.012449777788586087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,5120,0.01961599952644772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,512,0.020658666888872784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,2048,0.03237244486808777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,1536,0.00797155582242542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,10240,0.018208000395033095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,7168,0.03404711021317376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,7168,0.014789332946141561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,8192,0.004936888813972473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,8192,0.004932444542646408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,10240,0.08311111397213407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,512,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,7168,0.014003554979960123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,5120,0.021112889051437378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,16384,0.046012444628609546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,1536,0.010551999840471478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,128,0.005604444278611078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,7168,0.010743111371994019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,16384,0.030208001534144085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,6144,2.2551040649414062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,512,0.06881688700781928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,65536,0.25223199526468915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,3072,0.0076453329788313965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,65536,0.0839644471804301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,1024,0.06320799721611871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3072,128,0.009658666948477427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,768,0.005218666460778978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,8192,0.0099973330895106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,3072,0.007240888973077138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,12288,0.018935110833909776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,512,0.10432710912492539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,256,0.0028168888141711554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,1024,0.00462400002612008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,7168,0.010425777898894416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,10240,0.011706666813956367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,3584,0.7399271329243978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,7168,0.013443555268976422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,512,0.005252444495757421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,8192,0.04141955574353536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,128,0.005081777771313985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,3072,0.01536799967288971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,3584,0.014018666413095264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,8192,8192,0.01852888862291972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,5120,0.010001777774757808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,128,0.004968889057636261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,1536,0.005998222364319696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,7168,0.014175999495718213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,4096,0.042288889487584434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,1536,0.010429333481523724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,512,0.005958222266700532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,128,0.011733333269755045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,65536,0.061569776799943715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,7168,0.024832000335057575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,64,0.0028906667398081887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,64,0.001814222170246972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,5120,0.008432000047630733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,32,0.020248000820477802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,7168,0.03836089041497972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,65536,0.08601778083377415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,256,0.005201777650250329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,16384,0.01958044370015462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,1536,0.01901955571439531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,512,0.006684444430801604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,256,0.0052977779673205475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,65536,16384,0.1652782228257921
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,10240,0.21803911526997885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,6144,0.009010666774378883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,12288,0.013184000220563678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,64,0.002143111079931259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,16384,2.6559005313449435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,128,4096,0.005960000058015187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,512,0.014167111780908374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,256,0.0021546667234765156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,7168,0.014501333236694336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,8192,0.013723555538389417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,768,0.00544888898730278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,2048,0.022425777382320825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,8192,0.02678577767478095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,10240,0.02158222264713711
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,8192,0.013452444639470843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,16384,0.049903111325369946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,10240,0.04929600159327189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,32,0.002132444539003902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,8192,0.06857422325346205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,3072,0.009185777770148383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,12288,0.004864888886610667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,32,0.008676444490750631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,10240,0.025965332984924316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,4096,0.02643377747800615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,1024,0.0035626664757728577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,7168,0.005214222189452913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,3072,12288,0.030202666918436687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,10240,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,2560,0.004872000051869286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,7168,0.024295111497243244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,3072,0.038140445947647095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,128,0.023931556277804907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,65536,0.053865777121649853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,2560,0.008225777910815345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,128,0.004949333353175057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,2048,0.006267555471923616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,768,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,1536,0.02405244443151686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,4096,0.007939555578761632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,256,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,6144,0.051516443490982056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,12288,0.03327644533581204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,768,0.005325333525737126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,5120,0.023416888382699754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,65536,0.3209884431627062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,1024,0.01517955627706316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,128,0.004926222066084544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1536,2048,0.007282666862010956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,6144,0.143394668896993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,2048,0.006700444552633498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,256,0.002529777793420686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,5120,7168,0.013423999978436364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,2560,0.014138667119873894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,12288,0.02097333305411869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,128,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,7168,0.009995555712117089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,65536,0.1240124437544081
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,768,0.017850667238235474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,768,0.015535111228624979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,3584,0.007617777420414819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,2048,0.0063768890168931745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,2560,0.015185778339703878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,768,0.006625777731339137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,4096,0.009690666364298927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,3072,0.024489776955710515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,2048,0.005608888963858287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,65536,0.31882667541503906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,1024,0.018806222412321303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,3072,0.011192888849311404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,10240,0.022659555077552795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,16384,0.005541333307822545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,2560,0.014863111906581454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,5120,0.18627733654446074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,64,0.0021973333011070886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,512,0.007329777710967594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,4096,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,1536,0.015432889262835184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,3072,0.017709333035680983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,16384,0.04411111275355021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,4096,0.008105778031879002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,65536,0.22063467237684461
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,4096,0.008624888956546783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,3584,0.004960000101062986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,65536,0.06868800189759997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,1024,0.007282666862010956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,10240,0.01296533313062456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,1024,0.005682666682534748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,1536,0.008887999587588841
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,6144,0.027638221780459087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,10240,0.030010667112138536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,12288,0.011513777905040316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,65536,0.3716835445827908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,12288,0.04258222381273905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,256,0.008539555801285638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,2560,0.006638222270541721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,1024,0.005608888963858287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,3584,0.00526577772365676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,12288,0.08532710870107015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,1024,0.005943999936183293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,1536,0.015278221832381355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,2560,0.006953777538405524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,12288,0.009661333428488838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,512,0.008028444316652086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,256,0.004907555464241239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,5120,1536,0.008476444416575963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,5120,0.004566222015354368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,32,0.009986666341622671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,4096,0.02552000019285414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,512,0.015122666954994202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,12288,0.024898666474554274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,32,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,3072,0.03428533342149522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,32,0.024345777100986902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,10240,0.014153778553009033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,1536,0.01276622215906779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,512,0.006937777830494775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,512,0.020398222737842135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,3072,0.0069679998689227635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,6144,0.009218666288587783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,16384,0.09595377577675714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,65536,0.06157777706782023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,32,0.0051893335249688895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,16384,0.0742808911535475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,6144,0.011635555161370171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,2048,0.020264888803164165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,3072,0.02696711156103346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,3072,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,768,0.007278222176763747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,512,0.015928000211715698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,7168,0.027223111854659185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,4096,0.06317066484027438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,768,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,8192,0.023108444280094568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,3072,0.0072266666425599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,768,0.0028204443968004654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,4096,0.03335377905103896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,10240,0.021204445097181533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,1024,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,5120,0.015639111399650574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,2048,0.006207110981146495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,128,0.006970666348934174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,7168,0.012295111185974546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,7168,0.12915289402008057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,12288,0.0224924451775021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,7168,0.017121778594122995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,4096,0.03995555639266968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,6144,0.014071111877759298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,5120,0.01772533357143402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,768,0.005882666756709416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,2048,0.008647999829716152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,1536,0.0239991115199195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,2048,0.00675022229552269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,128,0.01734133395883772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,65536,0.08327822552786933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,2560,0.027066666218969557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,256,0.017895110779338412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,16384,0.04684533344374763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,128,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,8192,0.19002755482991537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,64,0.004871110949251386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,3584,0.03290044599109226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,5120,0.0053075556125905775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,3072,0.08214310805002849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,2560,0.020689777202076383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,3072,0.007286222444640265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,8192,0.040834665298461914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,16384,0.022827555735905964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,6144,0.005960888746711943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,65536,0.052836444642808705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,32,256,0.0064533332155810455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,64,0.007954667011896769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,3072,0.0074480002125104266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,32,0.028765334023369685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,12288,0.03947110970815023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1024,2560,0.00489777781897121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,768,0.005626666463083691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,768,0.01650399963061015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,768,0.007639110916190677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,128,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,1536,0.008430221842394935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,3072,0.006421333385838403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,5120,0.018181333939234417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,65536,2048,0.07242933246824476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,5120,0.02001333402262794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,65536,2.176696989271376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,65536,0.24526490105523002
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,6144,0.012104888757069906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,512,0.013800889253616333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,10240,0.16248355971442327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,5120,0.10500088665220474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,768,0.005738666488064661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,3584,0.016511110795868766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,6144,0.011039999624093374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,8192,0.006894222150246303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,6144,0.013424889081054263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,7168,0.14248266485002306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,3072,0.008232889076073965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2560,5120,0.021648888786633808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,6144,0.009682666924264696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,7168,0.013782222237851886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,64,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,16384,0.02910311023394267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,1024,0.014101333088344999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,65536,0.12517778078715006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,10240,0.011576889289749993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,12288,0.041181332535213895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,768,0.0028568889117903183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,6144,0.04610844453175863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,1536,0.005928888916969299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,128,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,6144,0.01725955473052131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,768,0.006656000183688269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,768,0.007770666645632849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,6144,0.009311111436949836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,16384,0.06692889001634386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,32,0.0021191111041439907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,8192,0.013688888814714221
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,64,0.0017955555684036678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,5120,0.02711822258101569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,1024,0.0064088888466358185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,1024,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,6144,0.01353422221210268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,7168,0.025610667135980394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,8192,0.028584000137117174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,2560,0.029682666063308716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,1536,0.0031519999934567344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,512,0.020403555697864957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,4096,0.008010666403505538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,3584,0.004934222333961063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,2560,0.009665778113736046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,4096,0.011958222422334882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,512,0.006624889042642381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,5120,0.013598221871587964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,512,0.011064888702498542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,5120,0.008730666504965888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,768,0.0028080000645584534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,5120,3072,0.02536444365978241
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,1024,0.020297777321603563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,3584,0.011708444191349877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,768,0.05647289090686374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,2560,0.005304000029961268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3584,256,0.005930666708283954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,2048,0.009023110899660323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,256,0.0021804444905784395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,8192,0.005250666704442766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,5120,3072,0.0083155557513237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,1024,0.03810311026043362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,3072,0.007287999822033777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,1536,0.04304977920320299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,8192,0.019267555740144517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,6144,0.03134311238924662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,6144,0.02238222294383579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,4096,8192,0.024152888192070857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,6144,16384,0.02460266649723053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,6144,0.027496889233589172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,65536,0.20432178179423013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,1024,0.015839111473825242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,256,0.05870755513509115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,2048,0.006586666736337874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,32,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,12288,2048,0.011698666546079846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,10240,0.0196142229768965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,256,0.030344890223609075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,768,0.006055111272467508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,2560,0.006797333144479328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,16384,0.04304800099796719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,128,0.014136888914638095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,16384,0.05262488789028592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,16384,0.02956355611483256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,7168,0.013329777452680798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,3072,0.02684266699684991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,512,5120,0.009394666386975182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,5120,0.013439111411571503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,32,0.005592888842026393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,4096,0.04685689012209574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,512,0.016515556308958266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,3584,0.007286222444640265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,6144,0.02537333303027683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1024,1536,0.00629066675901413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,5120,0.010122666756312052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,2048,0.004592000196377437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,2048,0.015290665957662793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,12288,0.036885334385765925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,3072,0.025960000024901494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,4096,0.01758399936887953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,256,0.0022026666750510535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,5120,0.028547555208206177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,5120,5120,0.03512977891498142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,768,0.017145777742067974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,4096,0.014764444695578681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,3584,0.007252444823582967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,3584,0.023004444109068975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,8192,0.010716444088353051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,256,0.0023093333260880574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,16384,0.041469332244661115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,10240,0.056032001972198486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2560,10240,0.015252444479200574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,128,0.013817778064144982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,65536,1.3920115364922419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,65536,0.1532284418741862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,2560,0.062090668413374156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,1536,0.0079644446571668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,5120,0.016528889536857605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,2048,0.017608000172509086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,768,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,32,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,32,0.0021786666992637846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,16384,0.020290666156344943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,1024,0.00794933322403166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,256,0.011708444191349877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,2048,0.0042035554846127825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,768,0.010670222342014313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,7168,0.026766222384240892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,2560,0.021035555336210463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,2560,0.0677493347062005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,4096,0.010344888601038191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,12288,1024,0.009762666291660732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,16384,0.043975998957951866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,8192,0.01035733355416192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,2560,0.08384178082148235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,2560,0.020958221620983545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,8192,0.0328435566690233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,4096,0.01001422190003925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,6144,0.17955288622114393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,256,0.01367111090156767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,256,768,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,5120,0.08080622222688463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,6144,0.02161955502298143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,768,0.015486222174432544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,65536,0.08773689137564765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,256,0.016515556308958266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,16384,12288,0.040103111002180315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,5120,0.008295111358165741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,4096,0.01477866702609592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3584,1024,0.003504000190231535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,65536,7168,0.14993511305914983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1536,7168,0.011053333679835001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,512,0.015184889237085978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,2560,0.011721777419249216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,1536,0.003569777641031477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,1024,0.007308444215191736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,512,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,2048,0.01107022249036365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,2048,0.015450666348139444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,10240,0.004606222112973531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3584,8192,0.014382221632533602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,32,0.0028613333900769553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,3072,0.007259555160999298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,32,10240,0.011369778050316704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,6144,65536,0.20013066132863364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,8192,5120,0.3613244427575006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,5120,0.006893333461549547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,10240,0.019276445110638935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,256,0.002469333302643564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,6144,0.010763555765151978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,1024,0.008655110994974772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,64,0.002475555572244856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,16384,0.12759021917978922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,3072,0.0069448889957533936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,7168,0.029789331886503432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,10240,0.06699733601676093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,65536,0.08975822395748562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,4096,0.04491288794411553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,16384,0.04461955693033007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3072,1024,0.005580444302823808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,4096,10240,0.03870755434036255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,7168,3584,0.024400000770886738
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,4096,0.018923555811246235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,1024,0.01512888901763492
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,2560,0.02958044409751892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2048,6144,0.018953777021831937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,256,0.017895999881956313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,12288,0.31999556223551434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,7168,0.023042667243215773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,256,0.0058808889653947614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,768,0.0029075555503368378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,512,0.012686221963829465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,12288,0.0351475543446011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,128,0.005902222047249476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,12288,0.2056293355094062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,1024,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,5120,0.01109155515829722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,5120,0.019221333993805777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,7168,0.02075466679202186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,2560,0.018882667024930317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,128,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,8192,0.028064890040291682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,3072,0.007283555964628856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,6144,0.009647111097971598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,64,0.0058986664646201665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,32,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,512,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,3584,0.006271111054552927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,10240,7168,0.043423109584384494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,5120,0.00867733359336853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,4096,0.00886311133702596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,65536,0.07818133301205106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,10240,0.07849599917729695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,6144,0.045017778873443604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,12288,0.012488000094890594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,512,0.005199999858935674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,128,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,4096,0.012835555606418185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,64,5120,0.008642666869693333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,1024,0.012824888858530255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,2048,0.009132444030708736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,1024,0.018581334087583754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,64,0.002157333410448498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,6144,0.03488533364401923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,32,1536,0.006049777898523543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,5120,0.07312711079915364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,512,0.006810666786299811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,16384,0.04281599985228645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,768,0.002699555622206794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,768,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,8192,0.010673777924643623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,7168,128,0.0041893331540955436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,10240,0.016460445192125108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,12288,0.01979022224744161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,128,0.001830222292078866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,3072,0.017657778329319425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,5120,0.01553422212600708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,3584,0.005941333456171884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,2560,0.004600000050332812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,16384,0.38252533806694883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,1024,0.006949333681000604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,2048,0.011295999917719098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,3072,0.01034577770365609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,65536,0.09128089083565606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,1536,0.01702133317788442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,2560,0.016885333591037326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,65536,32,0.01647911138004727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,256,0.011128889189826118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,3584,0.007310222420427535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,128,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1024,16384,0.020850666695170935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,3072,0.009986666341622671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,1536,0.007999111380841997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,3072,0.02459377712673611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,128,0.005633777628342311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,16384,10240,0.0539591113726298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,2560,0.004864000197913912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,1024,0.018166222506099276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,16384,16384,0.0721857746442159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,256,0.0021804444905784395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,768,0.0038951109680864546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,65536,0.07583644655015734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,512,0.0144533332851198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,1024,0.006266666783226862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,7168,0.01904355486234029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,6144,0.07150933477613661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,512,0.005234666582610872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,12288,0.020317332612143624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,7168,0.055348445971806846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,16384,0.07586933506859674
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,65536,6144,0.04291199975543552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,3072,0.013157332936922709
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,1024,0.003773333297835456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2048,12288,0.051536887884140015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,5120,0.004594666676388847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,16384,0.012491555677519904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1536,3584,0.007281777759393056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,10240,0.025010666913456384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,5120,0.1210106611251831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,65536,0.0972435540623135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,8192,0.03664266732003953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,256,0.014498665928840637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,1536,0.013479111095269522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1024,16384,0.01941777765750885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,8192,768,0.0032008888406885993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,12288,0.023975999818907842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,12288,0.04117244482040405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,512,0.06760088602701823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,5120,0.00831911133395301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,3072,0.01202488856183158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,10240,256,0.011996444728639392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,32,0.004934222333961063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,512,0.0069066666894488865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,2048,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,7168,0.027593778239356145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,12288,0.004560889055331548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,16384,0.011189333266682096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,6144,0.009311111436949836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,16384,0.11220978366004096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,12288,0.044887112246619336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,64,0.006303999986913469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,2560,0.04606311188803779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,2048,0.02776355544726054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,64,0.001855111163523462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,512,0.009703111317422655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,3072,0.007629333270920648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1536,16384,0.030877331892649334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,16384,0.017887110511461895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,3072,0.015182221929232279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,64,0.020273778173658583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,512,0.002509333400262727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,65536,0.16020533773634169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,32,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,5120,0.018627555833922494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,10240,0.01923822197649214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,2048,0.004184000194072723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,3584,0.01923199991385142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,32,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,768,0.0292231109407213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,3584,3072,0.027461333407296076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1536,3584,0.007624000310897827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,6144,0.009333333207501305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,64,0.0018088887963030075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,128,0.014106666048367819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,8192,0.014533332652515836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,10240,0.04502933224042257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,12288,0.06599999798668756
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,128,0.004873777843183941
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,64,5120,0.008285333712895712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,256,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,1024,0.010626667075686984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,32,12288,0.02437066700723436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,768,512,0.01517688896920946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,512,0.0053119998839166425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,3584,0.019851555426915485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,128,0.006319111006127463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,12288,0.5175475544399685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,16384,5120,0.04494933287302653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,2560,0.022447999980714586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,10240,0.018439110782411363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,1024,0.019578667150603402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,65536,0.12903466489579943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,768,0.02032800018787384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,65536,0.18449778027004668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,12288,0.06701955530378553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,10240,0.025099555651346844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,5120,0.10022577974531387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,7168,0.01568533314598931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,65536,12288,0.23919823434617785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,10240,0.018578666779730055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,2560,0.007281777759393056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,768,0.0024906665914588505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,1536,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,6144,0.018193778064515855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,128,0.015487111277050443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,10240,0.017159110969967313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3072,3072,0.009664889011118147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,256,0.004960000101062986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,3584,0.010109333528412713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,6144,0.024099555280473497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,128,0.005192888693677055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,3072,0.0181048893266254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,12288,10240,0.727874649895562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,256,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,32,0.0018586667461527719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,32,8192,0.010728889041476779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,8192,0.03927377859751383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,65536,0.2173173295127021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,32,6144,0.009321778184837764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,12288,0.03921866748068068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,2560,0.018553778529167175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,768,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,1536,0.006261333409282897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,8192,0.11216178205278184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,1536,0.030849777989917334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,1536,0.024079999989933435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,12288,0.01700711084736718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,65536,32,0.025238222546047632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,8192,0.044996443721983165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,128,0.00554666668176651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,1536,0.005956444475385878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,4096,0.00794666674402025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,65536,0.38573066393534344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,6144,0.009328888522254096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,1024,0.006267555471923616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,1536,0.01588444411754608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,10240,0.09420266416337754
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,64,0.0052622221410274506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,3584,0.014016888207859464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,12288,0.028047111299302842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,128,0.005615111026499007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,512,0.0028417776856157514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,2560,0.00867377801073922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,3584,0.010669333239396414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,65536,0.23220356305440268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,32,0.009977777798970541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,768,0.005888000130653381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,3072,0.0041982221106688184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,2048,10240,0.17537066671583387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,1024,0.007038222418891058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,256,0.004876444323195351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,2560,0.009120889008045197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,65536,0.137807117568122
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,3072,0.013737777868906656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,65536,0.05433600147565206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,5120,128,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,16384,0.006940444310506185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,3584,0.007643555601437886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,10240,0.10098755359649658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,2560,0.012647999657524956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,12288,0.026689777771631878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,10240,0.00574222207069397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,16384,1.2502258088853624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,1024,0.011604444848166572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,5120,0.011712888876597086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,10240,0.026017778449588354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,10240,0.10553955369525486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,1536,0.003535110917356279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,65536,1.969869401719835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,5120,0.013295110729005603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,7168,0.009656888743241629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,512,0.0053048887186580235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,3072,0.00721777809990777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,7168,0.006962666908899943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,128,0.006294222341643439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,3584,0.02162666618824005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,5120,7168,0.01570311188697815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,5120,0.004926222066084544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,256,0.00590133335855272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,32,0.0018399999373488957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,16384,128,0.002206222257680363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,10240,0.013148444394270578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,10240,65536,0.09490844276216294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,3072,0.009408888717492422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2560,65536,0.077711111969418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,2048,0.01699555582470364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,2560,0.0325271122985416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,2560,0.014532445205582513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,512,7168,0.00980444418059455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,2560,32,0.002116444417172008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,8192,0.018607111440764535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,10240,0.017472889688279893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,8192,0.010888888604111142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,10240,0.028706666496064927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,512,0.014456000592973499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,3584,0.19786577754550508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,256,0.0032071111102898917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,5120,0.011404444773991903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,3072,0.009166222479608323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,3072,0.04645511176851061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,2048,0.015099555253982544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,3584,0.015802666544914246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,65536,0.025123554799291823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,4096,0.028871999846564397
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,1024,256,0.005547555370463266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,1536,0.005895111295912001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,3072,0.053640888796912305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,64,8192,0.06596799691518147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,16384,0.02606311109330919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,512,0.002817777709828483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2048,8192,0.021366222037209406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,65536,0.13699821631113687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,2048,0.024897777371936377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,768,2048,0.006606222026877933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,2560,0.007618666523032718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,8192,0.04256177941958109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,2048,0.012362666428089142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,10240,0.037026666932635836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,3072,0.007361777954631382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,3584,0.028723554478751287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,5120,0.009000889129108852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,8192,0.014969777729776172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,2560,0.02443555494149526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,512,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,1024,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,768,0.002516444358560774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,8192,0.03186044428083632
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,256,768,0.006213333457708359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,3584,0.006623999940024481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,7168,0.013473778135246701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,128,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,8192,0.02160533269246419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,512,0.005202666752868229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,128,0.030615998639000788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,256,0.024472888973024156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,3072,0.010401777923107147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,128,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,4096,0.011333333121405708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,6144,0.13509244389004177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,768,5120,0.009287111461162567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,16384,0.04784889022509257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,6144,0.025594666600227356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,768,0.005541333307822545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,10240,0.015090665883488126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1536,65536,0.9698880513509115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2560,2560,0.007291555404663086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,512,0.0029093333416514923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,16384,0.12377866109212239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,6144,0.013408888545301227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,4096,0.008325333396593729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,65536,3584,0.02900000082121955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,128,0.0038951109680864546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,6144,8192,0.015433778365453085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,8192,0.08524533112843831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,7168,0.0177208897140291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,3584,0.007232888705200619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,16384,0.0421688887808058
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,5120,0.02643377747800615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,256,0.01441333360142178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,6144,8192,0.015836444165971544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,2048,0.012039110892348819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,7168,1536,0.020374221934212577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,2048,0.006274666637182236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,7168,8192,0.022123555342356365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,12288,0.012788444757461548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,3072,0.00980000032318963
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,1536,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1536,128,0.014111111561457316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,1536,0.0038559999730851916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2048,1024,0.0058755555914507965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,5120,0.03794399897257487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,512,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,64,32,0.002136888810329967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,16384,0.015116444892353482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,256,0.018937778141763475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,3584,0.013348444468445249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,1536,0.005577777822812398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3072,7168,0.035511109564039446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,512,4096,0.006533333410819371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,12288,0.013415111435784234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,4096,0.02050755586889055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,12288,0.013983999689420065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,768,0.01581333412064446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,16384,0.7831938001844617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,256,0.00488800017370118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,3584,0.01240799989965227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,64,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,2560,0.011030221978823343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,256,0.014137778017255994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,256,0.006248888870080312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,8192,128,0.05280266536606682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,3072,0.012078222301271228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,2560,0.007096000015735626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,65536,0.08668355809317695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,768,0.0801102254125807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,65536,0.5026328828599718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,32,1024,0.002850666642189026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,128,0.033583111233181424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,768,0.0032302221904198327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,128,0.0021822222818930945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,16384,0.015052444405025907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,10240,0.022294221652878657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,7168,0.038274665673573814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,3584,0.03817777832349142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,16384,0.02367111047108968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,512,0.035785777701271906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,128,0.002493333278430833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,512,0.0069164443347189165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,3584,0.0883786678314209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,2048,0.010535111029942831
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,128,0.02199911077817281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3584,1024,0.015449777245521545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,10240,0.017575111654069688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,512,0.00591822216908137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,128,5120,0.008955555657545725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,256,0.016864889197879367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,5120,0.011776000261306763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,64,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,3584,0.007419555551475949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,12288,0.011974222130245633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,768,0.014295111099878946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,768,0.04899466700024075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,1536,0.16941600375705293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,12288,0.022615111536449854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,256,0.0159608887301551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,1024,0.0069431112044387394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,128,0.00489688871635331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,1536,0.013412444127930535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,3584,0.03031022349993388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,12288,0.2927653259701199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,128,0.006400888992680444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,12288,0.049377779165903725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,6144,0.015498666299713982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,2048,0.015867556134859722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,4096,0.013893332746293811
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,16384,0.37679645750257706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,128,0.0028408887899584244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,16384,0.017873777283562552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,5120,0.020185778538386028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,5120,0.024737776981459722
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,7168,1536,0.006057777752478917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,6144,0.03291822142071194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,64,0.003544888976547453
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,512,0.014098667436175875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,32,0.011513777905040316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,32,0.0068488890926043195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,2048,0.007694222033023834
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3072,5120,0.008574222524960836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,2560,0.020233778489960563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,2560,0.004967999955018361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,1024,0.0032017777363459268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,3584,0.028958221276601154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,6144,0.005984888722499211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,2560,0.05786933501561483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,10240,0.01700444519519806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,7168,2560,0.007250666618347168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,2560,0.009354666703277165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,1024,0.07442222038904826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,10240,0.07290222247441609
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,4096,0.008020444048775567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,2048,0.006589333216349284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,7168,0.013263111313184103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,10240,0.011334222224023608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,4096,0.1632373332977295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,3584,0.03908088803291321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,64,0.013410666750537025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2560,16384,0.02407466702991062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,64,0.005239999956554837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,7168,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,3072,0.006933333145247565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,1536,0.006529777828190062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,65536,10240,0.0700862209002177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,128,0.021271111236678228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,2560,0.013286222186353473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,1536,0.008323555191357931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,1536,0.008974221845467886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,32,0.0017991111510329777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,3584,0.01740444368786282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,768,0.013801777528391944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,7168,0.013800000150998434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,8192,0.021967111362351313
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,6144,1536,0.04674488968319363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,2560,0.007287999822033777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2560,256,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,8192,0.017703110973040264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,10240,7168,0.058641778098212353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,2560,0.016864889197879367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,65536,0.046329776446024575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,8192,0.04117866688304477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,256,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,65536,0.1276142199834188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,128,0.005642666584915585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,1024,0.007271111011505127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,64,0.0018622221218215094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3072,16384,0.023963555693626404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,5120,0.026312000221676294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,16384,2048,0.009097778134875828
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,12288,0.020624000165197585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,12288,0.0654239985677931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,8192,0.03239911132388645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,768,0.0034977777136696708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,1024,0.011697777443461947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,64,0.005921777751710679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,768,32,0.006267555471923616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,65536,0.06756977902518378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2560,768,0.006948444578382704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,256,1024,0.011720000041855706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,6144,10240,0.019514666663275827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,16384,0.03284710976812575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,256,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,512,0.008638222184446123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,4096,0.016885333591037326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,2048,0.007870222131411234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,32,0.0021395554973019496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,512,0.0062773335311147906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,512,0.011416888899273343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,8192,0.0079297779334916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,65536,0.07288177808125813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,64,0.005852444304360284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,1024,0.012617777619096967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,768,0.017903111047214933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,12288,0.1012808879216512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,256,0.006551110910044775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,6144,0.011510222322411008
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,5120,0.007376888559924231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,8192,0.03021777669588725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,4096,0.011061333119869232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,2560,0.00628355559375551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,3072,0.03104355600145128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,10240,0.011732444167137146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,4096,0.016332444217469957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,128,0.006721777634488211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,64,0.0023084444304307303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,16384,0.01778133379088508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,512,1024,0.005994666781690385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,128,0.005626666463083691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2048,65536,0.07648533582687378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,4096,16384,0.021898667017618816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,12288,0.012167111039161682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3584,256,0.023136888941129048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,1536,0.008375999828179678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,8192,0.04649244414435493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,65536,0.11874400244818793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,512,0.005256888767083486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,32,0.0018542222678661346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,3072,0.007615110940403408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,4096,0.2237351205613878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,10240,0.4269706673092312
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,128,16384,0.011038222246699862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,3584,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,512,0.00489688871635331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,256,0.0052488889131281115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,128,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,65536,0.08006222380532159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,5120,65536,0.07567911015616523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,7168,0.020618667205174763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,768,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,4096,0.03856089048915439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,128,0.004892444445027245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,6144,0.014427555931939019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,6144,0.04182844360669454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,3584,0.015761777758598328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3072,7168,0.010958222051461538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,7168,0.01032266683048672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,64,0.007654222349325816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,12288,0.013111111190583972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,6144,0.011904888682895236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,1536,0.014785778191354541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,16384,0.02370044423474206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,4096,0.018519111805491976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,1536,0.06379110945595635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,1024,0.014449778530332776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,16384,0.02308711078431871
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,5120,0.008440000315507254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3584,256,0.002487999904486868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,7168,0.01070844464831882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,8192,0.04044622182846069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,512,0.017997332745128207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,128,0.00488711107108328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,768,0.013433777623706393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,7168,2048,0.028267555766635474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,65536,0.0877608921792772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,768,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,768,5120,0.009544000029563904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,5120,0.058628446525997586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,256,0.009212444225947062
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,8192,0.01239377756913503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,7168,0.020907556017239887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,16384,10240,0.212682671017117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,2560,0.007603555917739868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,7168,0.029380444023344252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,2048,0.016198222835858662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,256,0.002519999941190084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,10240,0.011392888923486074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,65536,0.05662844578425089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,10240,256,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,12288,0.07925599813461304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,3072,0.010642666783597736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,16384,0.05592266718546549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,1024,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,65536,0.06235910786522759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,16384,0.005884444548024072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,3584,0.014858666393491956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,768,0.008651555412345463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,4096,0.00794577764140235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,16384,128,0.002185777864522404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,5120,0.05813688702053494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,5120,0.013460444079505073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,16384,0.023004444109068975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,10240,0.02306488818592495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,65536,0.14326400227016872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,65536,2.2127289242214627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,12288,0.056314667065938316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,2560,0.009902222288979424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1536,2560,0.014834667245546976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,4096,0.009315555294354757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,512,0.004239999999602635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,256,0.0028657778683635923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,64,0.005223999834722943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,2048,0.006613333192136552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,5120,0.01180622229973475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,6144,0.012290666500727335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,65536,0.06538666619194879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,512,3584,0.007287999822033777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,64,0.002187555655837059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,2560,0.007280000381999546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,1536,0.0053191110491752625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,65536,0.01830577850341797
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,32,6144,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,4096,0.0185199992524253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,10240,0.012097777591811286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,64,0.006640888750553131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,3072,0.009647111097971598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,4096,0.010215111076831818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,4096,0.013415999710559845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,5120,0.026888888743188646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,128,0.0052897776994440295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,512,0.0052675555149714155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,10240,10240,0.47089423073662656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,7168,0.019335110982259113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,16384,12288,0.04126755396525065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,65536,0.12329422103034125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,2048,0.007283555964628856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,512,0.003992889076471329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,10240,0.02972266740269131
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,5120,0.01862844493654039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,65536,0.0788773364490933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,32,128,0.00796977761718962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,6144,0.016812443733215332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,12288,0.01995111174053616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,6144,0.01754133403301239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,8192,128,0.014363555444611443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,128,0.004946666873163647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,768,0.01701155636045668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,6144,0.020592000749376085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,5120,0.12143022484249538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,5120,2048,0.05216977662510342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,1536,0.025677333275477093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,2048,0.008640888664457533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,2560,0.004548444516128964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,8192,8192,0.04573066698180305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,65536,7168,0.1080097754796346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,65536,0.15225244892968073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,10240,0.012085333466529846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,768,0.00351555562681622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,3584,0.007387555307812161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,2048,0.01755644381046295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,8192,0.07215111123190986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,7168,0.10567466417948405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,2560,0.006631999793979857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,65536,0.013298666311634911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,128,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,128,3072,0.007234666910436418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,64,0.002543111021320025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,7168,0.029672000143263075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,512,0.008937777744399177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,1024,0.0035093331502543557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,4096,0.031565331750445895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,12288,0.02057422200838725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,5120,0.008651555412345463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,4096,0.01684533390733931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,128,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,768,0.0028719999310043124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,16384,0.053800887531704374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,2560,0.00654755574133661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,128,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,1536,0.006224888894293044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,2560,0.01975555552376641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,512,0.026579555537965562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,768,0.005554666535721884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,3584,0.5089919832017686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,32,0.005995555470387141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2560,7168,0.02372177773051792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,10240,0.0151235560576121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,8192,0.07991466919581096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,1536,0.004556444370084339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,65536,0.1633333365122477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,12288,0.1879671149783664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,3584,0.03776177763938904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,2048,0.017143110434214275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,3584,0.015231110983424716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,4096,0.006330666856633292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,1024,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,3072,0.011207111179828644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,7168,0.052130666044023305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,7168,0.0045422220395671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,12288,0.01683200067943997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,32,6144,0.009309333231714036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,128,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,5120,0.14687111642625597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,128,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1024,1536,0.0038417776425679526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,3584,0.025368889172871906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,1024,0.013807999591032663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,768,0.005960000058015187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,1024,0.0032213332338465583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,128,0.002141333288616604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,4096,0.008664000365469191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,8192,0.010695999695195092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,3072,0.028792887926101685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,16384,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,512,0.0035235554807715942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,1024,0.005903111149867375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,3072,0.009972444838947719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,65536,0.07518488830990262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1536,768,0.004206222378545337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,3072,0.010416888528399997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,6144,0.04253244400024414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,2048,0.007254222200976477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,1536,65536,0.5559457672966851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1536,4096,0.017041777571042378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,4096,0.009015999734401703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,6144,0.1183484395345052
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,6144,0.015442666080262927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,1024,0.03669155637423197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,1024,0.025424000289705064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,10240,0.30023733774820965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,6144,128,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2560,8192,0.012451555165979596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,2560,0.007376000285148621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,5120,0.03161777721510993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3584,64,0.013585777746306526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,128,0.00554577757914861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,12288,0.22074666288163927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,7168,0.6123893525865343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,3584,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,7168,0.02973955538537767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,8192,0.03016888764169481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,64,0.024347555306222703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,32,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,7168,0.027640889088312786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,2560,0.0063484443558586975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,2048,0.004173333446184794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,4096,0.00870311094654931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,128,0.005574222240183089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3072,7168,0.03860888878504435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,65536,0.06305333640840319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,10240,0.03368977705637614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,3072,8192,0.02825688984658983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,7168,0.013235555754767524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,16384,0.03960977660285102
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,7168,768,0.007609777980380588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,10240,0.041401777002546526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,7168,0.009964444571071202
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,3072,0.01629066632853614
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,512,0.025813332862324182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,8192,0.013767110804716745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,7168,0.026334222820070054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,7168,0.1616142193476359
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,8192,0.014208889669842191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,1536,0.00590133335855272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,65536,128,0.009015999734401703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,8192,10240,0.05078044533729553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,1536,0.00600888869828648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,16384,0.024005333582560223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,10240,0.015451555450757345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,65536,0.2459191216362847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,128,0.004918222212129169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,1024,0.018239999810854595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,65536,0.07752799987792969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,1536,0.006086222413513396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1024,2048,0.006583999842405319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,256,0.02199288871553209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,4096,0.05563910802205404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3584,256,0.0022293333378103045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,16384,0.024988444315062627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,5120,0.01751733322938283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,32,0.005210666606823604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,65536,0.6322542296515571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,12288,0.017671111557218764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,768,768,0.015428443749745687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,6144,0.08114222023222181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,8192,0.016020443704393175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,6144,0.012331555287043253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,65536,0.3418737782372369
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,5120,0.008640000389681922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,64,0.005467555589146084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,2560,0.024334222078323364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,2560,0.029558221499125164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,1536,0.0059057776298787855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,16384,0.026610665851169165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,65536,0.15154311392042372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,1536,0.005868444426192178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,4096,0.04629866613282097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,12288,0.026035555534892615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3584,10240,0.01661066710948944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,8192,0.005603555589914322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,16384,0.017401778035693698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,3072,0.007301333049933116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,65536,8192,0.17291288905673555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,7168,0.02225422196918064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,128,0.014128888646761576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,128,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,65536,0.1501377820968628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,5120,2560,0.00794044468137953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,5120,0.011963555382357704
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,16384,0.02123733361562093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,10240,0.02750933335887061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,6144,0.01591911084122128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2048,65536,0.07613066832224528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,768,0.01369866645998425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,128,0.0021626665774318907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,6144,0.034182220697402954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,1536,0.0320720010333591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,3072,0.025468443830808003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2560,512,0.012780444489585029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,2560,0.007270221908887227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,2048,0.007851555943489075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,7168,0.009657777845859528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,8192,3072,0.00998133338159985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,4096,12288,0.057573331726921924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,768,0.018533334136009216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,5120,4096,0.01943555474281311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,768,5120,0.008712000317043727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,8192,0.014117333624098035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,512,0.0155102229780621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,7168,0.013353777428468069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,10240,0.01888799998495314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,256,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,16384,0.08510844575034247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,12288,0.03935555617014567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3584,3584,0.010404444403118556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,7168,1024,0.0069288888739215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,1536,0.006594666590293248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,3584,0.036955555280049644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,2560,0.006969778074158563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,1536,0.006234666539563074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,512,512,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,10240,512,0.01851377718978458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,6144,0.015837333268589444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,1536,0.006259555617968242
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,10240,0.03605688942803277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,1024,0.06478222211201985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,1024,0.01310755560795466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,3584,0.0534746679994795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,2048,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,1024,0.011715555356608497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,1024,0.0069191112286514705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,768,0.005377777748637729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,12288,0.014084445105658637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,768,0.005294222384691238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,3072,0.005943111247486538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,10240,0.023364444573720295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,256,0.004249777644872665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,65536,0.17849600315093994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,16384,1536,0.03276711040072971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,10240,5120,0.02051911089155409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,2560,0.007644444704055786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,3072,0.024344000551435683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,2048,0.008634666601816813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,768,1024,0.007659555309348636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,64,8192,0.03865955604447259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,10240,0.01239111108912362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,8192,0.02253155575858222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,768,2560,0.020229332976871066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,5120,0.008275555239783393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,2560,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,12288,0.02938400043381585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,8192,0.011738667057620155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,768,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,5120,3584,0.009242666264375051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,1536,0.009934222532643212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,1024,0.005614222337802251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,7168,0.011416888899273343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,128,0.007596444752481248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,6144,0.017417778571446735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,8192,0.013167999684810638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,4096,0.03068888849682278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,2048,0.005578666511509154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,512,0.04651822315322029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,6144,12288,0.021262221866183813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,12288,0.023363555471102398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,16384,768,0.006275555739800136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,512,0.03987733191914029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,6144,0.010078222387366826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,5120,0.008397333323955536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,4096,0.0199635558658176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,3584,0.013454222016864352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,65536,0.10294133424758911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,64,32,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,4096,12288,0.018401778406567044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,3584,0.008671999805503422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,10240,128,0.0052666668262746595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,10240,0.06117866436640421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,16384,0.010710222025712332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2560,6144,0.009694221946928237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,768,0.016541333662139047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,2560,0.008645333349704742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,2048,0.006341333190600078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,65536,0.1573973364300198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,3584,0.011301333705584208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,8192,7168,0.014153778553009033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,5120,0.03186755710177951
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,512,0.012459555433856117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,2560,0.013013333082199097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,1024,0.007606222397751278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,10240,0.034973333279291786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,256,0.015978667471143935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,7168,0.014899555179807874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,4096,0.02472177810139126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,65536,0.793556425306532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,128,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,1536,0.006248000181383557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,1536,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,512,0.03982133335537381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,5120,0.040587554375330605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,6144,0.008982222113344405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,2048,0.004318222403526306
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,1024,0.002986666642957263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,6144,0.009170666337013245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,6144,10240,0.018619555566045973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,1536,0.008416000339719985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,5120,0.008871999879678091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,64,0.002918222298224767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,512,0.013810666898886362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,1536,0.0262773334980011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,2560,0.02951288885540432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,128,0.011391999820868174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,1024,0.005591111050711737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,2560,0.012310222619109683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,16384,0.06265422370698717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,12288,0.03280177712440491
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,4096,0.02200711104604933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,8192,0.01309866706530253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,6144,0.01076533314254549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,3584,0.007297777467303806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,12288,0.03787822193569607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,6144,0.06489155689875285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,10240,1024,0.020979555116759405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,768,0.005624889085690181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,6144,0.010727110836240979
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,2048,0.01089599976936976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,64,3584,0.007652444144090016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,10240,0.014475555883513557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,768,8192,0.009015111459626092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,5120,8192,0.014911111858155994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,1536,0.010604444477293225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1024,128,0.0021377777059872947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,2048,0.01553422212600708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,6144,0.04188977678616842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,7168,0.17306133111317953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,64,0.002161777681774563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,2048,0.02026844521363576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,2048,0.016233777006467182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,1536,0.0038808890514903595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,1024,0.012538666526476542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,128,0.004923555586073133
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,3072,0.021075555019908484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,2048,0.004231111043029361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,32,2048,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,768,0.017460443907313876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,8192,0.010774222513039907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,256,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,6144,1024,0.005050666630268097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,3584,0.009322666459613377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,16384,0.04211022125350105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,3584,0.011720889144473605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,1024,0.011007111105653973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,10240,65536,0.0944053332010905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,768,0.015185778339703878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,12288,0.013082666529549493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,10240,1.1992408964369032
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,8192,0.17971021599239775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1536,256,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,6144,0.010732444624106089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,3072,0.00795555528667238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,1024,0.014555555250909595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,2560,0.13052799966600206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,4096,0.007635555333561367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,5120,0.03678044345643785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,3584,0.012799999780125089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,2560,0.004911111046870549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,512,0.003528000166018804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,4096,0.006622222148709827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,6144,0.01219555570019616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,1024,0.0031848889258172777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,16384,0.044864889648225575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,4096,0.08727733294169109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,3072,0.012848888834317526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,8192,0.01332622187005149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,768,0.016021332807011075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,7168,0.024431111084090337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,16384,0.20775377750396729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,8192,0.030184000730514526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,7168,0.023013333479563396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,8192,0.035587555832333036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,512,0.11711733871036106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,65536,1536,0.06650133265389337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,64,5120,0.004527111020353106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,512,0.009356444080670675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,1536,0.010208889014191097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,7168,0.010659555594126383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,1536,0.005966222120655908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,7168,16384,0.043440000878440015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,1024,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,5120,0.034279998805787824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,3072,0.0041955556306574075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,128,0.0038968887594011095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,4096,0.04185511006249321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2048,2560,0.018209778600268893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,8192,0.02650755478276147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,16384,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,64,65536,0.1019866665204366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,32,0.007663999994595845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,2048,0.00489155575633049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,512,1536,0.005950222412745158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,4096,0.029186666011810303
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,3072,0.010057777994208867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,768,7168,0.024751999311976965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,1536,0.029498666524887085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,6144,64,0.0018035556293196147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,2560,0.00490044429898262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,6144,0.02719733284579383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,10240,0.012099555797047086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,128,0.005247999810510212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3584,64,0.011076444553004371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,128,0.006245333287451003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,2048,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,12288,0.038768000072903104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,512,0.003175111073586676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,3072,0.010802666346232096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3584,2048,0.006575111299753189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,128,0.003850666599141227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,2560,0.006918222126033571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,6144,0.010686222049925061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,512,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,512,0.0027822222974565295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,768,0.006122666514582104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,64,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,4096,768,0.005888000130653381
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,4096,0.011404444773991903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,128,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,64,0.0052791109515561005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,1536,0.006601777755551868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,12288,3584,0.010694222317801582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,6144,0.043007999658584595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,2048,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,6144,0.01548266741964552
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,65536,0.640150229136149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,128,512,0.013478221992651621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,5120,0.005585777676767773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,128,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,6144,0.017587555779351127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,65536,0.09778400262196858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,2560,0.0069271110826068455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,3072,0.009700444837411245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,4096,0.010037333601050908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,10240,0.016490666402710807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,4096,0.007652444144090016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,7168,0.017535999417304993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,256,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,10240,0.032235556178622775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,512,0.017522666189405654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,7168,0.1800995535320706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,4096,0.00489688871635331
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1536,1536,0.017469333277808297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,6144,0.013631111217869652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,32,0.0038764443662431505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,2560,0.008639111287064022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,5120,0.009316444396972656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,7168,0.0099973330895106
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,32,0.005896889087226655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,7168,0.011385777758227454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,10240,0.017482666505707633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,768,0.005570666657553778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,32,0.005457777943876054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2048,3072,0.015456888410780164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,6144,0.04588355620702108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,12288,0.017887999614079792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,7168,2560,0.07046933306588067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,32,0.005231110999981563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,16384,4096,0.012961777547995249
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,3072,0.006933333145247565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,2560,0.012749333348539142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,8192,0.011861333416567909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,256,16384,0.016819554898473952
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,3072,0.015810666812790763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,6144,0.007849777738253275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,16384,0.04541422261132134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,12288,0.020288000504175823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,2560,0.027585777971479628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,256,0.006295111030340195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,2560,0.007585778004593319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,16384,0.1436648898654514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,2560,0.007296000089910295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,10240,0.017676444517241586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,10240,0.03851911094453599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,5120,512,0.04081511166360643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,3072,0.008970666262838576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,10240,0.015977778368526034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,32,0.002209777840309673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,2560,0.004852444347408083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2048,32,0.007670222057236566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,3072,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,256,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,2048,0.013436444103717804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,65536,2560,0.12228177653418647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,512,0.012611555556456247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,16384,0.25385067198011607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,6144,0.008987555901209513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,256,0.004565333326657613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,10240,0.015810666812790763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,1536,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,768,0.00590133335855272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,8192,2560,0.006968888971540663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,512,0.005254222287072076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,12288,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3584,256,0.006913777854707506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,10240,0.012853333519564735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,3072,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,7168,0.030829333596759375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,2048,0.006255110932721033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,32,0.005247999810510212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,5120,0.008369777765538957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,1024,0.01802399920092689
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,2560,0.006937777830494775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,4096,0.008635555704434713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,3072,0.008956444760163626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,65536,0.026915555198987324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,256,768,0.005576888720194499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,256,0.0028088889602157804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,7168,0.01703644461101956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,256,12288,0.01401511165830824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,10240,0.016585777203241985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,4096,0.029567112525304157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,6144,16384,0.05868888563579983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,7168,0.0052426668504873914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,1536,0.004552888787455029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,8192,128,0.013444444371594323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,512,65536,0.07512088616689046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,1024,0.006950221955776215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,3072,128,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,256,0.030140442980660334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,10240,0.011687999798191918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,2560,0.008604444563388824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,2048,0.020630222227838304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,2048,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,6144,0.010062221851613786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,1536,0.021347555849287245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,12288,512,0.007285333342022366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,512,0.005913777897755305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,64,2048,0.006608888920810487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,16384,0.017228444417317707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,512,0.006677333265542984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,512,0.0052915554907586836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,2048,0.015778667396969266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,10240,0.02051911089155409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,12288,256,0.006590222318967183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,2560,0.007236444287829929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,65536,0.2682017750210232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,8192,0.017602667212486267
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,3072,0.008736888567606607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,2048,0.029901332325405542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,10240,0.014544000228246054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,5120,0.013165333204799227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,768,0.0035084444615576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,8192,0.02442399991883172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,7168,256,0.005615111026499007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,5120,0.010464888479974536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,16384,0.1967688931359185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,8192,0.021112889051437378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,768,0.002846222163902389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1024,16384,0.02232355541653103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,5120,0.03086755673090617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,16384,0.3268488778008355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,6144,0.01037688884470198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,65536,0.04731377628114489
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,32,0.010096888575288985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,16384,3072,0.011069333387745751
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,3072,0.021917333205540974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,256,0.005893333504597346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,3584,0.026527111728986103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,2560,0.0074720001882976955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1536,3584,0.0075235557225015424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,16384,2560,0.03049955434269375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,7168,0.023988443944189284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,7168,0.013384000294738345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,512,0.013834666874673633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,8192,2048,0.006623111251327727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,768,0.003843555433882607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,128,0.0052968888647026485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,2048,0.006606222026877933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,128,0.014139556222491793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,512,0.014831110835075378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,7168,0.013086222112178802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,768,0.01955111159218682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,4096,1024,0.013546666337384118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3072,6144,0.012219555675983429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,1024,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,1536,0.017252445220947266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,128,16384,0.016695999436908297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1024,2048,0.004528000122971005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,7168,0.020220445262061227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,12288,0.01868977811601427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,3072,0.011648000114493899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1536,12288,0.01591377788119846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,7168,0.010668444136778513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,10240,0.15388711293538412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,7168,0.021549334128697712
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,1536,0.01688799924320645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,2560,0.006931555353932911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,32,3072,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,10240,0.014671999547216626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,5120,0.01926577753490872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,2048,0.006907555378145642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,12288,0.030528889762030706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,128,2560,0.022291556000709534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,3072,0.008296888735559251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,128,0.005246222019195557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,8192,0.02614044480853611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,128,0.005564444594913059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,6144,0.01569600072171953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,128,65536,0.05151822169621786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,4096,5120,0.016202666693263583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,3072,0.03461066550678677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1024,128,0.014478221535682678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,2560,0.014144000079896716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,16384,0.1962613397174411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,16384,0.038952890369627215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,1024,0.002995555599530538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,65536,0.044320887989468045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,512,0.012910222013791403
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,768,3072,0.005482666608360078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,4096,65536,0.0679324467976888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,16384,65536,2.8110951317681203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,7168,0.016021332807011075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,16384,3584,0.04401955670780606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,768,0.021299555897712708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,768,7168,0.013252444565296173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,768,0.16693333784739176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,128,0.0021831111775504216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,5120,0.17877688672807482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,128,32,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,16384,1024,0.025387555360794067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,1536,0.01352355546421475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,6144,0.017859554953045316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,16384,0.02475111186504364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,12288,0.07491377989451091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,64,0.00510044437315729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,128,0.006632888896597757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,768,0.005571555760171678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,768,0.004685333205593957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,4096,0.01920622256067064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,2560,0.009764444496896531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,65536,12288,0.2419431209564209
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3584,3072,0.05159822106361389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3072,256,0.005640888793600931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2560,7168,0.025504888759719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,6144,0.05819911426968045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,6144,5120,0.01385600037044949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,512,0.01181422256761127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,4096,12288,0.02159999973244137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,6144,0.013785777820481194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,4096,0.006298666612969504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,1536,0.0035253332720862497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,3072,0.01577333278126187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,10240,0.013392888837390475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,3584,0.03410933415095011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1536,7168,0.0326186650329166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,3584,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,12288,0.021905778182877436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,6144,0.010425777898894416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,2048,12288,0.027064889669418335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,65536,0.07267822159661187
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,256,5120,0.009304888546466827
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,7168,0.014111111561457316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,128,0.0021502222451898786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,768,0.20763466093275282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,5120,0.011639110743999481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,2048,0.006582222051090664
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,1024,0.012094222009181976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,2048,0.04082311193148295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,16384,2560,0.018657777044508193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,8192,0.04747555653254191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,12288,0.01795644395881229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,12288,0.02642222245534261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3072,768,0.016192000773217943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,10240,3072,0.03725422090954251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2048,5120,0.027796443965699937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,32,0.0025519999778932994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,7168,0.015959110524919298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2560,2048,0.004269333349333869
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,2048,0.025690666503376428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,12288,0.02199999988079071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,3072,0.022815111610624526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,2048,0.00600355573826366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,65536,768,0.016851555969980028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,3584,0.007323555648326874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,7168,0.009721777505344814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,128,0.002187555655837059
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,16384,0.03234311276011997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,256,0.007252444823582967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,4096,3072,0.007257777783605788
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,768,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,10240,2048,0.021440888444582622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,7168,0.014076444837782117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,128,5120,0.009352888498041365
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,7168,0.013354666531085968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,12288,0.22201244036356607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,8192,768,0.011483555866612328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,8192,2560,0.04604444570011563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,12288,0.047832889689339526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3584,128,0.0148026661740409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,512,0.004190222256713443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,256,0.010729778144094678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,5120,0.01662400033738878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,2048,0.04930222365591261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,1024,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,1536,0.017892445127169292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,65536,5120,0.14815378189086914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,256,0.0021760000122918025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,512,0.0026826666047175727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,12288,32,0.0026888888743188647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,768,0.016491555505328707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,3072,0.013004444539546967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,1536,0.006640000061856375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,128,0.00554044461912579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,10240,0.02136444383197361
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,8192,0.0366977784368727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,768,0.0069724445541699724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,65536,0.08988710906770493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,32,0.0018622221218215094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1024,5120,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,10240,0.01204622205760744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3584,2048,0.018823110395007663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,768,0.016939555605252583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,6144,5120,0.019525332583321463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,512,7168,0.009687999884287516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,256,0.013798221945762634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,768,0.005604444278611078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,2560,0.008653333617581261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,1536,0.006656000183688269
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,6144,0.0511813329325782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,2560,1536,0.042620443635516696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1536,256,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,16384,0.005454222361246745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,512,128,0.0038542221817705366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,128,0.013754666679435305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,3584,0.00693866651919153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,10240,0.05275288886494107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,6144,0.010741333166758219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,128,0.0023866666273938283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,3072,0.022980444961124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,8192,0.007308444215191736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,512,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3584,256,0.0023235556566052968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,10240,0.040046221680111356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,6144,0.04171466827392578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,256,256,0.005271111097600725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,1024,0.005879111174080107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,3584,0.027521777484152053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,12288,0.05987022320429484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,2560,0.009737778041097853
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,128,0.014152000347773233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,128,6144,0.01178311142656538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,2560,0.004950222041871813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,1024,0.003170666595300039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,2048,0.0041831110914548235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,512,16384,0.02792177928818597
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,6144,0.03901866740650601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2048,3584,0.007999999655617608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,1536,0.006705777926577463
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,10240,0.21871466106838652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,3584,0.14713244967990452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,3584,0.08972266647550796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,1536,0.012804444465372296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,3584,0.025024889243973628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,6144,3072,0.007032000356250339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,6144,2048,0.021614222062958613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,512,5120,0.010662222074137794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,2560,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,12288,0.01886755559179518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,768,0.01035288886891471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,4096,0.009959110783206092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,2560,0.006648889018429651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,12288,128,0.0021840000732077495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,10240,8192,0.07825155390633477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,6144,5120,0.021758221917682227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,3584,0.016959110895792644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,2048,0.012089778151777057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,12288,1024,0.018810666269726224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,7168,0.010989333192507425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3072,128,0.01682400041156345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,5120,0.06855733527077569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,1024,0.033719112475713096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,1536,0.004866666677925321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,16384,0.018756444255510967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,65536,0.14713688691457114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,7168,0.01886488828394148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,128,0.016057777735922072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,3072,0.006956444846259222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1024,2560,0.01035911093155543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,65536,512,0.05502488878038195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,12288,0.048467556635538735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,7168,256,0.003132444495956103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,32,8192,0.010392888552612728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,8192,0.004550222307443619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,4096,0.024855999483002558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,7168,0.004573333180612988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,2048,0.013362666798962487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,64,0.0021742222209771476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,16384,10240,0.06608711348639594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,2048,0.004515555583768421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,256,0.014500444134076437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,5120,0.029509332444932725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,4096,0.050026665131251015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,10240,0.03917866614129808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,2560,0.021039111746682063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,5120,0.03401599989997016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,7168,0.10113866461647882
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,3072,0.007254222200976477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,7168,0.009700444837411245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2048,8192,0.030204445123672485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,5120,768,0.016528889536857605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,4096,6144,0.025612443685531616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,5120,0.007269333634111617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,12288,0.025040000677108765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,3584,0.007605333295133378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1536,1024,0.014087110757827759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,7168,16384,0.07263200150595771
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,1024,0.008988444175985124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,32,0.010336000058386061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,5120,0.02184711065557268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,7168,32,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,1024,0.014456000592973499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,1536,8192,0.05435199870003594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,16384,0.017516444126764934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,8192,0.014167111780908374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,65536,0.12466577688852946
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2560,10240,0.03202666507826911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,32,0.0021760000122918025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,5120,256,0.003232888877391815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,512,0.01812533371978336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,12288,4096,0.02956355611483256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,768,0.006222222414281633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2560,3584,0.008270222279760573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,65536,0.03190044562021891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,128,0.007920888562997183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,3584,0.011434666812419891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,5120,0.06301510996288724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,768,0.023753777146339417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,4096,0.01000711073478063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,12288,0.028016000986099243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,256,0.005933333188295364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,5120,0.03217333224084642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,3584,0.015201777219772339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,6144,256,0.01648266613483429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,32,1024,0.0028213332924577924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,7168,0.04312444395489163
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,32,0.0018453333112928604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,2560,2048,0.016864889197879367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,768,0.0036159998012913596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,7168,12288,0.020184889435768127
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,128,6144,0.008975110948085785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,128,32,0.0018151110659042995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,12288,12288,0.0537484453784095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,6144,0.014942222171359591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,12288,0.019903999235894945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3072,1536,0.006232000059551663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,12288,5120,0.038410666916105486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,2048,0.03219199842876858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,64,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,2560,0.006578666468461354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,1024,0.005611555443869696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,65536,0.058747556474473744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,5120,0.012392889294359418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,5120,0.025607110725508794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,8192,0.011429333024554782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,12288,0.01885688967174954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,768,0.005934222290913264
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,8192,0.03257511059443156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,5120,0.060452441374460854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,5120,0.009860444400045607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,768,2048,0.00904177791542477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,7168,0.043059554364946157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,65536,0.279200898276435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,8192,0.02060622142420875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,4096,0.007655111451943715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,8192,6144,0.027952889601389568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,65536,0.024209777514139812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,1536,0.0040835556056764396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,128,0.006609777609507243
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,128,0.014495111174053617
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,4096,0.008603555460770925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,6144,1536,0.008973333570692275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,3584,0.007328888608349695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,512,0.0053386667536364655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,65536,0.06743644343482123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,12288,0.029422223567962646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1024,8192,0.010355555348926121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,512,0.00629333323902554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,2560,0.014104000396198697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2048,12288,0.013005333642164866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,8192,3584,0.024296889702479046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,5120,0.020282665888468426
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,1536,0.045752889580196805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,3072,0.045758220884535045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,7168,0.031163556708229914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,7168,0.030264887544843886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,12288,0.035608000225490995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,512,0.0038480001191298165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,128,0.015973332855436537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,5120,0.023491554790072974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,16384,0.05003733436266581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,128,0.0021315556433465746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,12288,0.015244444211324057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,4096,0.013739555246300168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,3072,0.025380444195535447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,2048,0.017920000685585868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,4096,0.009146666361225976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,32,0.005973333285914526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,65536,0.4943857722812229
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,1536,0.006221333311663733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,128,10240,0.012097777591811286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,256,0.006167999986145232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,6144,0.014104000396198697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,6144,0.09189777904086643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,2048,5120,0.008656000097592672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,64,0.001855111163523462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,256,0.0021537778278191886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,5120,0.012805333567990197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,16384,0.024696888195143804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,6144,0.028099556763966877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,128,0.0052177777720822235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,1024,0.003334222154484855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,5120,256,0.004958222309748332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,12288,256,0.011401777466138205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,3584,3584,0.007662222617202335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,65536,0.013597333596812354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,2560,0.03288266725010342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,768,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,128,0.013767110804716745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,5120,1536,0.004308444344335132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,3072,0.014663999279340109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,65536,16384,0.1713013384077284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,6144,0.010024000373151567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,2560,0.023457777169015672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,6144,0.00928177767329746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,8192,0.02774311105410258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,10240,0.016813332835833233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3072,8192,0.0248364441924625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,256,0.015493333339691162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,4096,0.01443822185198466
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,10240,0.034607112407684326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,64,0.005530666559934616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,3072,768,0.01907822158601549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,10240,512,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,6144,0.03530222177505493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,3072,0.06378489070468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,8192,0.015843555331230164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,256,0.00517866677708096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,8192,0.019896888070636325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2560,256,0.004939555707905027
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,3584,0.02716800073782603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,4096,0.007646222081449296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,4096,0.005659555395444234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,768,0.014096889230940076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,5120,0.03323910964859857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,10240,0.028877334462271795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,8192,0.027997334798177082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,16384,256,0.0041537777417235905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,2560,0.01240711079703437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,4096,6144,0.03974044322967529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,2560,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,1024,65536,0.07660266425874498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,6144,256,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,65536,0.09301600191328262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,7168,2560,0.012234666281276278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,1536,0.009723555710580613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,768,0.003185777821474605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,5120,0.008928000099129146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,512,0.010023111270533668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,6144,0.008999110923873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,2048,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,5120,0.11906754970550537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,3584,0.014645333091417948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,64,0.006583999842405319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,256,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,512,0.0031644445326593188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,5120,0.0041751112374994485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,5120,0.0461768905321757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,128,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,256,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,256,0.011762667033407422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,8192,0.023994666006830003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,7168,0.01575644479857551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,12288,0.04756711257828606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,2048,0.00673333348499404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,128,2048,0.00629333323902554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,3072,0.007262222468852997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,512,0.006646222124497096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,10240,3072,0.012094222009181976
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2048,1024,0.005573333137565189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,2560,0.005247999810510212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,128,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,4096,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2560,512,0.0031679999083280563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,6144,3584,0.008013333711359236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,10240,0.23526310920715332
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,8192,3072,0.026561778452661302
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,7168,0.01578044394652049
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,8192,5120,0.014131555954615274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,16384,0.006607111129495833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,6144,0.11558577749464248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,5120,3584,0.026214222113291424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3072,768,0.005587555468082428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,64,0.0028293333533737394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,65536,0.35071643193562824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,65536,0.1100471085972256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,256,0.005937777873542573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,1536,0.0059368887709246735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,8192,0.02237866653336419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,2048,0.006308444258239534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,10240,0.013980444934633044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,6144,0.01684266659948561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,7168,0.021271111236678228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,128,6144,0.017535999417304993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,8192,0.013737777868906656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,65536,0.1595004399617513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,65536,0.44414401054382324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,64,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,12288,3584,0.015096889601813423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,65536,0.07580444547865126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,10240,0.04654133319854736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,2560,10240,0.02661422226164076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,4096,2560,0.006597333484225803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,12288,5120,0.038351999388800725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,6144,0.013958222336239286
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,256,8192,0.02993511160214742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,256,0.0024888888001441956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,16384,768,0.026407111022207472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,3072,0.01825777855184343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,32,0.007771555748250749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,12288,2048,0.025273778372340735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,16384,0.02590933276547326
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,10240,0.07532888650894165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,3072,0.01035288886891471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,6144,2560,0.025039111574490864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,32,0.0018168888572189545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1536,6144,0.009010666774378883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,256,0.019575110740131803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,768,7168,0.02478133307562934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,2048,4096,0.007612444460391998
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,1024,0.015429332852363586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,512,0.010724444356229572
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,16384,0.031176888280444678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,256,0.005565333283609814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,5120,0.022629333866967097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,8192,0.09882133536868626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,16384,0.272998227013482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,65536,10240,0.11132089296976726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,768,0.004955555415815777
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,6144,0.0148026661740409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,7168,0.0103502223889033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,3584,10240,0.04625155528386434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,512,0.02199111051029629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,3584,0.029116445117526587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,5120,7168,0.01312622262371911
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1024,1536,0.00832711077398724
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,8192,0.01276622215906779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,32,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,5120,16384,0.05720622009701199
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,5120,0.008351999852392409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,1536,0.006272888845867581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,16384,0.026050666968027752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,12288,0.023959111836221483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,3072,0.06861066818237305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,65536,7168,0.15437422858344182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,512,0.00249422217408816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,7168,0.03390044305059645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,1024,0.026963555150561865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,10240,0.01680799987581041
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,6144,0.018570666511853535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,16384,10240,0.034033778640959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,1024,0.058871997727288135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,65536,0.1483902268939548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,10240,0.0317831105656094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,64,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,4096,0.0926355587111579
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,256,0.007958222594526079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,2048,0.0047679998808436925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,768,0.0025395556456512874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,32,128,0.001828444500764211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,4096,5120,0.1364648871951633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,65536,0.07684355311923556
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,64,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,10240,0.042641778786977134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,3584,0.004237333519591225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,12288,128,0.01443555619981554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,3072,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,2560,0.007505777809354994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,5120,768,0.021037333541446265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,3072,0.005232000102599462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,3584,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,5120,768,0.016176000237464905
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,12288,0.05040711164474487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,768,0.015466666883892484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,256,0.005586666779385672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3072,1536,0.006599999964237213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,768,0.005533333453867171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,128,2560,0.00962933318482505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,8192,2560,0.010779555473062726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,256,0.005597333527273602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,128,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,256,0.002504000026318762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,64,1024,0.003107555624511507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,256,0.0123848890264829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,512,768,0.003149333306484752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1024,5120,0.0069360000391801195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,7168,0.017849778135617573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,2560,0.006602666858169768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,12288,0.019052444232834708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,256,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,768,0.04841511117087471
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,10240,0.011719110939237805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,3584,0.012372444073359171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,8192,768,0.006237333433495627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,768,0.018735110759735107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,768,0.0052328887912962176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,64,0.004872888740566042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,7168,3584,0.056925336519877114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,7168,0.019327999817000497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,7168,8192,0.01685599982738495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,10240,0.02162222233083513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,768,0.006312888943486744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,256,1024,0.0031662223239739737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,32,10240,0.02723377777470483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,64,0.004575110971927643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,2048,65536,0.09979822238286336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,5120,0.013845333622561561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,16384,0.03853244582811991
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,64,0.005306666509972678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,7168,2560,0.024894222617149353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,4096,0.007496889266702864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,512,8192,0.010712888505723743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,4096,0.007970666719807519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,10240,64,0.005910222315125995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,7168,10240,0.026319111386934917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,512,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,7168,0.05925955375035604
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,16384,0.05678311321470472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,4096,5120,0.023358222511079576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,256,0.007636444436179266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,512,0.0351822210682763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,65536,1536,0.3838355541229248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,16384,0.023054222265879314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,8192,0.5509093072679307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,8192,0.12179466088612874
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,16384,0.023026666707462732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,10240,0.01643733349111345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,65536,0.07906222343444824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,256,0.0021964444054497615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,256,0.007251555720965068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,3584,0.06778844197591145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,7168,10240,0.03338133295377096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,2560,0.005378666851255629
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,768,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,768,512,0.002893333426780171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,2560,0.01549600064754486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,1024,0.014935111006100973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,1536,0.013074666261672974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,32,1536,0.009862222605281407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,10240,768,0.006851555572615729
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,5120,0.01238222254647149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,7168,0.021624000536070928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2048,768,0.005281777845488654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,768,0.006261333409282897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,65536,2048,0.19686933358510336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,8192,10240,0.018269333574506972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,2560,0.010048000348938836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,6144,0.03150755498144362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,2048,65536,0.14176177978515625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,2048,0.004582222137186262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,3584,0.03036710951063368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,128,8192,0.005616888817813661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,2048,0.00452977791428566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,8192,0.012425777812798819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,1536,0.006181333214044571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,2560,0.03652711047066583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,256,0.03177777926127116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,3584,32,0.005946666830115848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3072,64,0.01924622224436866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,16384,10240,0.056259552637736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,4096,0.015871110889646742
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,10240,8192,0.025215999947653875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,512,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,4096,0.035312000248167254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,1536,0.005964444329341252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,3072,32,0.0064604443808396654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,1024,0.01752088963985443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,2048,0.006660444455014334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,3584,0.007626666790909237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,32,0.005245333330498801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,4096,0.2073662281036377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,6144,0.011056000159846412
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,256,0.0028213332924577924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,12288,5120,0.012960889273219638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,768,0.007280000381999546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,1024,0.012019555601808759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,2048,0.004306666553020477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,4096,0.05040533343950907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,8192,16384,0.027719999353090923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2560,12288,0.08006222380532159
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,4096,1536,0.005973333285914526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,2560,5120,0.00927466650803884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,3072,0.009454222189055549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,4096,0.007956444389290279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,16384,0.48498397403293186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,1024,0.008775999976529015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,3072,0.04832355512513054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,2048,16384,0.0240328891409768
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,16384,0.01681777834892273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,256,0.0052151112920708126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,12288,1536,0.010025777750545079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,10240,0.01761866609255473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,3584,0.039296888642840914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,2048,0.0064151109092765385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,4096,0.2917333443959554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,1024,0.004581333448489507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,6144,0.015840000576443143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,4096,8192,0.014669333895047506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1024,2048,0.006279110908508301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,6144,4096,0.025821333130200703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,512,0.0053084443012873335
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,3072,0.00794311116139094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,16384,0.018176888426144917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,6144,65536,0.08486755688985188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,3584,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,128,768,0.002801777794957161
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,3584,0.006627555522653792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,8192,0.018921777606010437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3584,10240,0.03159200151761373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,512,0.012118221984969245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,3072,0.0742151141166687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,6144,0.009326222042242685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,256,256,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,1024,0.003850666599141227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,32,0.0062871111763848206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,4096,0.030382222599453394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,1536,1536,0.005952888892756567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,3072,6144,0.009957333405812582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,12288,12288,0.06816266642676459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,10240,0.026975999275843304
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,16384,6144,0.02421244482199351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,7168,0.03421688742107815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,128,65536,0.02518044412136078
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,768,0.006247111078765657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,768,256,0.005778666585683823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,128,0.006958222223652734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,2560,0.01534577707449595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,1536,0.01959466603067186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,4096,8192,0.01399466726515028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,16384,16384,0.03338755501641168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,12288,0.01906488835811615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,32,0.00832088871134652
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,256,6144,0.01311466677321328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,1024,768,0.0053137776752312975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,16384,0.017163554827372234
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,7168,0.013192888763215808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,4096,6144,0.013015111287434896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,128,0.01445777714252472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,2048,0.006983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,4096,0.00795911086930169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,256,0.019368888603316415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,768,0.005551110953092575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,4096,0.007669332954618666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,768,128,0.0052062223354975385
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,768,256,0.004927999857399198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,512,0.012439111040698158
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,2560,0.049499554766549
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,12288,0.02366488840844896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,3072,0.03474844495455424
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,8192,0.0640266670121087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,1024,5120,0.018204443984561496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,10240,512,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,1536,0.007696888513035244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,1024,0.011071110765139261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,128,0.00552977787123786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,65536,0.04452266626887851
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,256,32,0.006235555642180973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,5120,0.02060622142420875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,3072,0.0627982219060262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1536,10240,0.014102222190962898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,1024,0.003177777760558658
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,512,8192,0.009986666341622671
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,12288,12288,0.049599998527103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,2560,0.00943288869327969
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,512,0.012437333663304647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,32,0.005940444353553984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,3584,0.0861919985877143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,256,0.002847111059559716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,128,0.005511111269394557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,12288,0.027330666780471802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2560,1536,0.03480799992879232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,6144,0.0059262220230367445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,5120,0.015417777829700045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2048,512,0.006264888991912206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,128,1536,0.006239999913507038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,64,3584,0.00490488898422983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,1536,256,0.002143111079931259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,8192,0.0144213338692983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,32,64,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,768,0.008674666285514832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,2048,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,1536,256,0.006565333240562015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,512,0.015523556205961438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,2560,0.007294221884674496
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,512,768,0.015551111764378019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,5120,3072,0.14746222231123182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,7168,0.028346667687098186
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,6144,0.013291555146376291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,3584,0.00795999997191959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,1024,1536,0.003617777592606015
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,6144,0.029740442832310993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,768,1536,0.013122667041089801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,768,0.004997333304749595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,1536,0.0196115556690428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,12288,256,0.01687022215790219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,32,0.0018417777286635505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,12288,0.012990222209029727
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,128,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,3072,0.007293333609898885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,5120,0.02106844385464986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,12288,0.016152888536453247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,2560,0.0075706665714581805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,6144,0.01745244363943736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,2048,0.008815111385451423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,6144,0.010001777774757808
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,4096,32,0.002159111201763153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,8192,0.010383110907342698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1024,768,0.023408000667889912
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,7168,0.042612443367640175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,5120,0.008667555948098501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,768,0.007319110963079665
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,256,0.004548444516128964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,16384,8192,0.07478044430414836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,7168,0.02193244463867611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,12288,0.020223110914230347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,768,0.002962666667169995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,4096,0.005613333235184352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,12288,0.03885599970817566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,6144,0.02381955583890279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,3072,0.0069822221994400024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,32,0.002191999927163124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,256,0.002519999941190084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,1024,0.005631999837027655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,4096,0.03351555599106683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,2560,0.007295110987292395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,2560,0.02389066749148899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,1024,0.005566222386227713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,128,0.0024951110697454875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,7168,0.01034044474363327
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,16384,0.010984000232484607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,32,0.0021955555097924336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,4096,256,0.014779556128713818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,3584,0.007267555428875818
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,7168,0.05015288790067037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,512,4096,0.011360888679822287
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,5120,0.008964444200197855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,128,0.0018515555808941524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,2560,0.0069146665434042616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,2048,0.024715556038750544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,768,16384,0.023951111568344965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,768,0.011572444604502784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,256,12288,0.00958577791849772
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,10240,0.020245333512624104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,128,3584,0.013768889009952545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,6144,0.02442488902144962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,65536,0.6751182344224719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,65536,2560,0.029322667254341975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,12288,0.013109332985348172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,2048,0.006976000136799282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,256,7168,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,3584,0.012991999586423239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,128,1024,0.006284444282452266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,768,128,0.002500444443689452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,512,65536,0.054526223076714404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,768,0.003124444435040156
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,12288,0.037520001331965126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,16384,0.1355608834160699
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,3584,512,0.00592533333433999
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,10240,0.010802666346232096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,1024,1024,0.012741333080662621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,32,0.0018479999982648427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,1024,0.0028684443483750024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,128,0.0052266667286554975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,16384,3072,0.025632888078689575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3584,256,0.004944889081848992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,1024,0.009896000226338705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,3584,0.10565155744552612
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,512,16384,0.1544462177488539
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,2560,0.01608000033431583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,2048,2048,0.006440889090299606
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,65536,1024,0.028055111567179363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,1024,0.017942221628295053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,4096,6144,0.012476444244384766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,8192,0.009139555195967356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,1024,0.0042364444169733255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,2560,0.01146133326821857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,10240,65536,0.09464889102511936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2048,6144,0.009679999616410997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,4096,0.0511084430747562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,5120,0.06983111302057902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,4096,0.024528000089857314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1024,256,0.002499555548032125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,256,0.016188444362746347
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,256,10240,0.03674044542842441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,1024,0.005970666805903117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,5120,65536,0.0820995569229126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,8192,0.020255110330051847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,768,0.016145777371194627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,512,0.013221333424250284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,8192,7168,0.03427911135885451
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,256,0.018930666976504855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,1024,0.022597332795461018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,5120,0.00830577810605367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,5120,0.014173333843549093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3072,128,0.013783111340469785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,65536,0.10974666807386611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,2560,0.008323555191357931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,768,0.013739555246300168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,256,3072,0.007596444752481248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,10240,5120,0.01833866702185737
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,64,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,65536,0.2827395598093669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3072,5120,0.010226666927337646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,256,0.00216977774269051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,8192,12288,0.03989333245489333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,3584,0.03975466556019253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,2048,0.008641777767075432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,1024,0.022053332792388067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,2560,512,0.013716444373130798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,5120,1024,0.0057964444988303715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,5120,0.15587022569444445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,768,3584,0.07020088699128893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,10240,0.01754577789041731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,10240,0.035932444863849215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,16384,0.02481777800454034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,5120,256,0.0159244438012441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,8192,0.03880177934964498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,4096,0.03298222356372409
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,10240,0.019288000133302476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,16384,0.4775102403428819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,6144,8192,0.016047999262809753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,16384,0.055496888028250806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,1536,0.0035297779573334586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,10240,0.045216888189315796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2048,12288,0.014816888504558139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,4096,65536,1.3082266913519967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,12288,7168,0.01625066664483812
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,16384,256,0.015455111861228943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,10240,0.14277156194051108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,10240,0.020258666740523446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,768,0.0144213338692983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,65536,0.07553866836759779
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,2048,4096,0.012424888710180918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,3072,0.008946666287051307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,256,0.014483556151390076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,10240,4096,0.015259555644459195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2560,7168,0.011386666860845355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,65536,7168,0.15501777331034342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,16384,6144,0.023551111419995625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,2560,0.006928000185224745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,16384,0.015498666299713982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,256,4096,0.015127999915017022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,7168,0.010303110712104374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,64,0.0021555556191338436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,3584,0.019721777902709115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3072,16384,0.022679110368092854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,128,0.003527111063400904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,7168,0.017532444662517972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,512,2048,0.013165333204799227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,1536,0.006134222365087933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2048,2048,0.020015110572179157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3072,128,0.008972444468074376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,128,768,0.017869333426157635
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,32,0.005473777651786804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,4096,0.00832444429397583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,1536,0.025137777129809063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,3072,0.007258666886223688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,12288,0.03948622279696994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,256,0.03773777683575948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,768,0.005506666584147348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,32,0.0018337778747081757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,12288,10240,0.17786755826738146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,4096,0.01220355596807268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,8192,256,0.006633777585294511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,1536,256,0.004921777794758479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,4096,0.01240355521440506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,512,0.006919999917348226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,2048,3072,0.0069217777086628815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,32,0.0018560000591807896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,256,0.016499555773205228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,10240,0.04722222354676989
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,32,0.004960000101062986
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2560,12288,0.013552889227867126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,64,128,0.006242666807439592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,4096,128,0.005550222264395819
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,5120,0.008608889248636033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,768,0.026777777406904433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,768,3584,0.007598222129874759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,3584,12288,0.019654222660594516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2560,6144,0.06219644678963555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,128,0.0038942222793896994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,5120,0.004544888933499654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,512,3584,0.006246222390068903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,512,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,8192,0.01643111142847273
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,3072,0.010777778095669217
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,12288,0.02202577723397149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,2048,0.004556444370084339
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,10240,0.05548266569773356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2048,7168,0.07442400190565321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1024,1536,0.003810666501522064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,8192,16384,0.05234044459131029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,16384,7168,0.042938666211234204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,3072,0.01344622257683012
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,32,5120,0.005542222410440445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,10240,4096,0.07393244240019056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,7168,64,0.001867555495765474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,128,0.005558222118351195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,512,12288,0.0439626673857371
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,7168,0.017380444539917838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,512,768,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,12288,3072,0.01536177761024899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2048,2560,0.017521777086787753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,3072,0.01889955500761668
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,8192,0.012310222619109683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,4096,10240,0.01623644431432088
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,7168,0.012800888882742988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,3072,0.09832800096935695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1536,16384,0.029523554775449965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,512,0.012072000238630505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,3072,4096,0.014064000712500678
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,7168,64,0.0025128889828920364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,5120,8192,0.06820444266001384
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,128,2048,0.008323555191357931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,12288,0.018586667047606576
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,12288,7168,0.03555466731389364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,3072,1536,0.01648177703221639
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,7168,1024,0.016843555702103507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,128,0.0021439999755885866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,768,256,0.005664000080691443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,768,256,0.005280888742870755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,2560,5120,0.008640000389681922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,2560,0.04981866810056898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,256,768,0.005209777918126848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,5120,0.014785778191354541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,4096,5120,0.010959111154079437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,1024,0.04468444320890638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,128,1024,0.0059004442559348206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,5120,0.041132443481021456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,10240,32,0.00646666685740153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,65536,0.056815114286210805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,128,0.013712888790501488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,10240,0.020616888999938965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,7168,12288,0.03780355718400743
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,7168,0.011081777513027191
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1536,1536,0.006199999815887875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,1536,0.00922400007645289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,128,0.014095111025704278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,64,0.00254488881263468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,5120,6144,0.012802666260136498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,7168,0.013481777575280933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,128,10240,0.012070222033394707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,768,768,0.006260444306664997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2048,64,0.008988444175985124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2048,65536,0.169995559586419
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,4096,0.008295999632941352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,6144,65536,0.08132533232371013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,16384,5120,0.014344000154071383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,8192,0.7771306567721896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,768,0.006937777830494775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1024,4096,0.009322666459613377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,10240,0.04276266694068909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,128,2048,0.014109333356221518
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,1024,0.023032888770103455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,16384,0.029190222422281902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,5120,0.016203555795881484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,6144,2560,0.01645599967903561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,256,0.00217333332531982
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,4096,0.00794044468137953
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,6144,0.005024888863166173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,512,0.009342222577995723
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3072,512,0.015103111664454142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,7168,3072,0.010688888529936472
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,10240,0.3477991157107883
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,5120,0.017280000779363845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,3584,0.08278755346934001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,5120,0.022115555074479844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,1536,6144,0.01797422269980113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,256,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,5120,0.007062222394678328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,32,0.006620444357395172
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,64,0.0018337778747081757
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,512,0.004907555464241239
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,16384,0.012992888689041138
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,3584,0.058647109402550586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,12288,0.01483022173245748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,16384,4096,0.02217688825395372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,1536,0.024493333366182115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,3584,128,0.013752000199423896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,7168,0.012349333200189801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,16384,128,0.006239111224810283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,4096,6144,0.020824000239372253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2560,128,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,12288,0.07002577516767713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,1536,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,5120,6144,0.014105778601434497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,128,32,0.0017964444640609955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,1024,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,10240,0.01550222271018558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,16384,3072,0.0140657772620519
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1536,3584,0.030214223596784804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,512,12288,0.02334844403796726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,65536,4096,0.6855795648362902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,12288,0.012119111087587146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3584,16384,0.023762666516833838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,8192,3072,0.028708444701300725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,768,0.01681422193845113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,2560,0.08117333385679457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,256,0.002480000050531493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1024,512,0.007849777738253275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,7168,0.02221155497762892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,128,0.0021466666625605687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,1536,0.02090311050415039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,8192,0.01069777790043089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,3072,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,768,2048,0.0041786668201287585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,1536,0.007275555696752336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,2048,0.0053119998839166425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,3072,0.03269422385427687
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,32,6144,0.005927111125654644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,7168,256,0.0031591111587153543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,65536,0.06422133578194512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,12288,0.017973333597183228
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,10240,0.040411555104785495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,6144,0.009107555780145857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,7168,1536,0.01962222158908844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,2048,0.014102222190962898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,3072,0.03157155712445577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,128,12288,0.02066044509410858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,768,0.005629333357016246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,2560,0.026927999324268762
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,10240,0.050712890095180936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,2560,0.053357332944869995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,8192,0.03724177678426107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,7168,0.012058667010731168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,65536,0.04918933245870802
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,128,1024,0.008676444490750631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,6144,0.008975110948085785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1024,5120,0.027757333384619817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,2560,0.1846888860066732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,10240,512,0.007618666523032718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,64,3584,0.009993777506881291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,256,3072,0.013433777623706393
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,128,1536,0.034252444903055825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2048,6144,0.010030222435792288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,5120,256,0.00554666668176651
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,32,0.0028204443968004654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1536,128,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,2560,0.009018667042255402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,3584,2560,0.008678221868144141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,8192,16384,0.026349332597520616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,3072,0.007198221981525421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,3584,0.007658667034573025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1024,2560,0.006721777634488211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,3584,0.007641777396202087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,512,0.0328764451874627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,16384,0.02420177724626329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,2560,3072,0.020631111330456205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,3072,0.0052595556610160405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2560,5120,0.008966222405433655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,1024,0.013272888958454132
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,768,7168,0.009954666925801171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,1024,5120,0.008329778081840938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,64,4096,0.007639110916190677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3584,512,0.0029004443850782183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,8192,0.01601777805222405
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,512,0.0027137777457634607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,1536,0.005951111101441913
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,1536,0.008340444829728868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,768,8192,0.020045333438449435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,10240,7168,0.015377778146002026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,4096,0.00796088907453749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,2048,0.05924266576766968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,2048,0.015814221567577787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,256,0.005049777941571342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,6144,0.015357333752844067
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,10240,3584,0.02942844563060337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,5120,0.06764266888300578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,16384,0.02397777802414364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,2048,0.022073777185546026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,256,2560,0.01386844449573093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3584,65536,0.07914933231141832
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,4096,0.03715022073851691
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,5120,10240,0.029979556798934937
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,5120,256,0.0058764442801475525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,768,0.0028239999794297745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,3584,0.0662062234348721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,16384,0.039140443007151283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,65536,0.10628266466988458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,7168,32,0.002542222125662698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,65536,0.0785928898387485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2048,12288,0.011406222151385413
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,12288,0.861381318834093
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,5120,0.011145778000354767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,10240,0.011525332927703857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,2048,0.004271999829345279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1536,7168,0.012148444851239523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,5120,5120,0.01274311128589842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,3072,0.010075555907355415
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2048,2560,0.004952888935804367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,256,0.005292444593376583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,256,0.006690666493442323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,4096,6144,0.014937778313954672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,12288,0.08751910924911499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,10240,0.031567109955681696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,3584,0.007251555720965068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,16384,0.013423999978436364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,12288,0.01796888808409373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,3072,16384,0.021359110871950786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,2560,256,0.005115555392371284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,8192,2048,0.07622400257322523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,32,3072,0.006962666908899943
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,7168,0.14709333578745523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,2560,0.019925332731670804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,1536,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,1536,0.023980443676312763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,6144,0.022506665852334764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,10240,0.017252445220947266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,65536,0.09010222223069932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,8192,0.038873778449164495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,512,128,0.004920000003443824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,768,65536,0.24539645512898764
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,4096,0.01073688848151101
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,12288,0.12158132923973931
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,1024,6144,0.01887822151184082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,12288,0.02162933349609375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,5120,0.028113779094484117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,12288,2560,0.04159199860360887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,5120,3072,0.03714488943417867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,5120,0.008918222453859117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,512,12288,0.02197066611713833
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,3584,0.007659555309348636
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,2048,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,256,0.014440889159838358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,4096,16384,0.04147555430730184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,128,32,0.006601777755551868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,128,256,0.004916444420814514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,4096,0.009002666506502364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,1536,5120,0.008383110993438298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,512,0.016541333662139047
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,128,64,0.0018515555808941524
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,128,256,0.006253333141406377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,3072,0.007881777981917063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,2048,0.00831822223133511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,10240,512,0.018597332967652213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,65536,0.14035199748145208
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,8192,0.020647999313142564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,4096,12288,0.050019555621676974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,1024,2560,0.01716444392999013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,768,6144,0.017252445220947266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,7168,256,0.005958222266700532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,768,65536,0.021624000536070928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,16384,0.00526311124364535
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,1536,0.005605333381228977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1536,2048,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3584,5120,0.029955555995305378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,6144,128,0.008991111483838823
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,5120,3584,0.008290666672918532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,128,0.014458666245142618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,1536,0.23630844222174752
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,7168,0.019854222734769184
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,65536,0.06980266835954455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,5120,12288,0.02972888946533203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,256,0.005294222384691238
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,128,0.014460444450378418
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,5120,0.006227555374304454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,1024,0.0058133333093590206
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,16384,0.010099555883142684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,8192,0.015402666396564908
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,1024,0.005607111172543631
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,8192,1536,0.018210666047202218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,32,2560,0.0069288888739215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,1024,0.002842666581273079
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,768,768,0.0028159999185138275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,65536,1024,0.012362666428089142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,2048,1536,0.006887110984987683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,6144,64,0.009979555176364051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,65536,512,0.0786000026596917
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,8192,0.024720888998773363
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2560,32,0.0027520000520679685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2560,1024,0.009677333136399588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,12288,0.055789331595102944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,10240,0.13829955789777967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,1024,0.011173333558771344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,1536,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,5120,0.008919110728634728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,256,0.009663999908500248
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,7168,0.14065333207448324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,16384,16384,0.049442665444480054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,512,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1024,1536,0.013422222601042854
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,8192,2560,0.024392889605628118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,12288,0.056361774603525795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,8192,64,0.023545778459972803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,8192,0.011383111278216044
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,1024,128,0.0022808888720141514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,2560,0.02407377792729272
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,1536,0.006682666639486949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,16384,0.04892266790072123
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1536,2048,0.006923555499977536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3584,128,0.0052737775776121355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,768,10240,0.008278222547637092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,2048,7168,0.008931555681758458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,65536,256,0.023728888895776536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,256,128,0.0052373334765434265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,2560,0.012808000048001608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,32,0.005655111124118169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,512,0.017861333158281114
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3072,768,0.006583999842405319
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,7168,0.012913777596420713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,4096,0.0107342220014996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,2560,128,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,8192,0.0320355561044481
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,3072,7168,0.012133333418104382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,4096,6144,0.01313866674900055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,128,1024,0.005952888892756567
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,10240,1536,0.009392889009581672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,12288,0.012168000141779581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,32,512,0.002516444358560774
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,3584,0.010372444159454769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,6144,2048,0.038416001531812884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,3072,0.007306666837798224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,3584,0.00830844458606508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,7168,10240,0.01906488835811615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,64,0.00999111102686988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,256,0.01686755485004849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,16384,3072,0.21977333227793375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,4096,0.015003555350833468
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,6144,3072,0.17157067192925346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,1536,0.004535111288229625
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,1024,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,3072,0.01737333337465922
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,32,0.005961777849329843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,2048,0.01071111112833023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,768,0.0069715554515520734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,12288,0.4836631351047092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,65536,0.1895893282360501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,2048,0.07508266634411283
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,1024,128,0.004942222187916438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,3072,512,0.0028577778074476454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,7168,16384,0.17858755588531494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,3072,0.012427555190192329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,12288,0.024393777052561443
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,12288,1024,0.011350221931934357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,7168,3072,0.00998577806684706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,1536,7168,0.02502577834659153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,2048,0.011392888923486074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,10240,12288,0.06560711065928142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,2560,1536,0.01686755485004849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,2048,0.0052871112194326185
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,8192,0.02030311193731096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,10240,0.031054221921496924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,4096,128,0.013930666777822705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,10240,16384,0.04078311059210036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,16384,8192,0.7045840157402886
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,256,8192,0.011030221978823343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,16384,0.0221760008070204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,768,3584,0.012421333127551608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3584,4096,0.008652444514963362
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1536,16384,0.01908266709910499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2048,512,0.02859555681546529
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,8192,0.0867919988102383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,512,3072,0.006945778098371293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,5120,16384,0.04160088962978787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1536,3584,0.020982222424613103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,512,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,4096,10240,0.1610142257478502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,256,512,0.01899111105336083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,6144,0.020615110794703167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,3584,0.10886310868793064
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,2560,2048,0.019720888800091214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,32,2560,0.009393778112199571
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,512,1536,0.013196444345845116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,12288,6144,0.02740977704524994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,65536,768,0.2384586599138048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,12288,5120,0.013191999660597907
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,512,0.010847999817795224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,4096,768,0.02676711148685879
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,32,0.0028284444577164124
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,8192,2048,0.00664533343580034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,6144,128,0.00795911086930169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,2560,3072,0.006281777802440856
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,64,64,0.0018364443547195857
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,6144,0.027962668074501887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,2048,0.006276444428496891
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,5120,4096,0.028318223026063707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,768,0.003909333298603694
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,3072,16384,0.08466400040520562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,2048,0.006604444649484422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,256,768,0.006628444625271692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,16384,0.04456622070736355
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,12288,0.039224889543321394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,256,0.031160000297758315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,64,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,12288,256,0.005219555563396878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,8192,10240,0.027632888820436265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,2560,0.05184977915551928
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,5120,0.010501333408885531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,32,0.005233777893914117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3072,3072,0.017817777064111497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,65536,128,0.01956622302532196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,3072,0.026865777042176988
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,4096,2048,0.016993777619467843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,512,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,768,0.014821334017647637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,3072,4096,0.007875555919276344
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,5120,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,16384,1536,0.029218667083316382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,3584,65536,0.1525919967227512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,512,0.0028799999919202593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,12288,10240,0.02970577942000495
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,12288,768,0.005575110928879843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,2048,0.008616000413894653
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3584,768,0.005220444252093633
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,1536,0.006244444598754247
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,8192,0.023694222172101338
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,8192,0.01092088884777493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,3584,0.043099555704328746
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,3072,0.01442577772670322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,12288,0.0794666674402025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,4096,6144,0.022468444373872545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,65536,2048,0.20031021700965035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3584,32,0.002491555487116178
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,1024,0.006240889016124938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,256,1536,0.003843555433882607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,8192,16384,0.028922667105992634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,2048,0.046262221203909985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2560,1024,0.003863999827040566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,6144,0.0352222224076589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,3072,0.007992000215583378
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,7168,6144,0.014371555712487964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,10240,6144,0.03573777940538194
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,4096,128,0.0021484444538752237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,1536,0.006215111249023014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2048,65536,0.05905066596137153
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,7168,256,0.016364443633291457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,3584,0.007531555162535773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,2048,0.007601777712504069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,8192,0.08710133367114598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,3584,0.01571555601225959
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,512,8192,0.012381333443853589
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,32,1536,0.0041662222809261745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,2048,0.026406221919589575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,16384,512,0.006588444527652528
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,8192,1024,0.00795199970404307
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,8192,4096,0.010671999719407825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,768,7168,0.05482844511667887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,128,2048,0.007903999752468532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,2048,0.006311111152172089
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,5120,512,0.010218666659461128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,5120,7168,0.014346667461925082
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,7168,32,0.002473777780930201
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,10240,512,0.009358222285906473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,8192,0.027782221635182697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,3072,0.0072266666425599
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,8192,0.016546666622161865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,3072,0.013418667018413544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,6144,0.11669600009918213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,7168,128,0.019560888409614563
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,7168,6144,0.014707555373509726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3072,1024,0.004948444250557157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,128,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,128,0.004548444516128964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,1024,0.08422933022181193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,3072,65536,0.09333333041932847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,6144,5120,0.013949332965744866
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,8192,0.024895111719767254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,512,0.0025333333760499954
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1536,3584,0.0808782246377733
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,10240,0.01164177805185318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,512,0.0028400001012616684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,3584,0.005559111220969095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,256,1024,0.0028906667398081887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,2048,8192,0.02609688871436649
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,10240,0.017521777086787753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,6144,0.03709866603215536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,16384,2560,0.010667555862002902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,128,0.011706666813956367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,64,64,0.0019137778629859288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,4096,0.019476445184813607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,4096,0.007999111380841997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,12288,0.011728000309732227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,2048,0.006249777972698212
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,65536,0.07368977864583333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,6144,512,0.005201777650250329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,6144,6144,0.018391110830836825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3072,1536,0.014008000493049622
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,32,2048,0.004562666846646203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,4096,5120,0.01534577707449595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,16384,1536,0.012426666915416718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,128,0.0024968888610601425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,7168,768,0.05775022506713867
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,256,0.00281866660548581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,6144,0.013284444808959961
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,65536,0.218304885758294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,3584,2048,0.019295111298561096
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,6144,3584,0.009568000005351173
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,8192,0.010355555348926121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,512,0.00792711145348019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,16384,4096,0.017154667112562392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,7168,0.009979555176364051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,1536,0.00591822216908137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,256,16384,0.005985777825117111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,5120,0.007315555380450354
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,768,128,0.005620444400442972
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,4096,64,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,4096,0.01792266633775499
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,256,0.014117333624098035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,1024,3584,0.022592888938056097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,256,0.005877333382765452
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,8192,0.09878310892317031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3584,3584,0.010390222072601318
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,128,768,0.005572444448868434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,2560,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,7168,512,0.0052355556852287715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,1536,0.005535999933878581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3072,6144,0.16153155432807073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,512,7168,0.009789333575301701
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,4096,0.005953777581453323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3072,5120,0.008330666356616551
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,256,512,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,4096,0.007661333514584436
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,7168,6144,0.036031110419167414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,768,0.00555377784702513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,4096,0.018260444204012554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,1536,0.00795911086930169
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,8192,0.15934844811757407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1024,8192,0.008639111287064022
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,2560,0.009628444082207149
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,3584,0.007270221908887227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,64,0.00488711107108328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,3584,128,0.01442488862408532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,64,16384,0.01513066722287072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,32,0.011747555600272285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1024,10240,0.032531556155946516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,12288,8192,0.037951111793518066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,2560,32,0.007592000067234039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3072,8192,0.015063110325071545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,6144,8192,0.020607999629444547
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,256,0.002481777841846148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,65536,128,0.02431022293037838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,2048,0.006406222366624408
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,512,64,0.006721777634488211
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,256,0.009002666506502364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,32,64,0.0052151112920708126
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,3584,3584,0.008247111406591203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,256,4096,0.00794933322403166
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,1536,1024,0.017842666970358956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,2048,0.07926311095555623
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,10240,16384,0.062021334966023765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,7168,0.03795377744568719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,10240,0.05676800012588501
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,6144,2560,0.009340444372759925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,16384,0.026398221651713055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3072,16384,0.02033333314789666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,32,2048,0.014825777875052558
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,5120,0.008634666601816813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,16384,0.032122668292787336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,4096,0.01455200049612257
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,128,0.005230222311284807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3072,16384,0.03292355603641934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,4096,0.015512888630231222
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,768,0.00554933316177792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3584,3072,0.007601777712504069
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,1536,0.01239733315176434
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,768,6144,0.009305777649084726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,5120,0.00869422240389718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,256,128,0.006296000132958095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,1024,0.027104000250498455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,512,0.05513688921928406
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,8192,5120,0.030761778354644775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,4096,3072,0.009330666727489894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,4096,1024,0.015146666102939181
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,6144,0.027618666489919026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,7168,0.015787555111779105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,256,768,0.00519466648499171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,256,7168,0.015799111790127225
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,256,0.004899555610285865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,5120,0.008300444318188561
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,16384,10240,0.03564533260133531
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,10240,3072,0.025384000606007043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,7168,0.02017866737312741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,4096,4096,0.009591999981138442
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,65536,0.07553688685099284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2048,3584,0.023397333092159692
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,128,512,0.004959110998445087
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,16384,0.04113511244455973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,128,512,0.004906666775544484
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,65536,64,0.004937777916590373
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1024,5120,0.00868977771864997
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,4096,4096,0.02846222122510274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,10240,5120,0.4582986831665039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2048,3584,0.012069332930776807
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,10240,0.04694844285647074
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,256,1024,0.015584000282817416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,3584,1536,0.016932444439993966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,32,0.0018497777895794974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,2048,0.04261422157287598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,32,0.014171555638313293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,6144,3584,0.010943999720944298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,3584,65536,1.7483253479003906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,6144,1536,0.019926221834288705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,3584,0.007654222349325816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,5120,16384,0.035748445325427584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1024,6144,0.01761244402991401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,512,7168,0.02027466727627648
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,256,8192,0.028858668274349634
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,1024,0.026183999247021143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,256,64,0.005312888986534542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,2560,0.006979555719428592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,2048,0.004252444538805219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,16384,8192,0.03487199876043532
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,2048,0.02161600026819441
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,7168,0.014782221780882942
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,5120,10240,0.029178665743933782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,10240,768,0.007771555748250749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,12288,512,0.019564444820086162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,512,0.005286222116814719
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,1024,16384,0.016827555166350473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,16384,768,0.0188400000333786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,12288,2560,0.019665777683258057
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,1024,0.02528711160024007
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,3584,0.019680000013775296
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,12288,1.0784888797336154
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,5120,8192,0.018391999933454726
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,32,0.007288888924651676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,6144,0.09990311331219143
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,768,0.023373333944214716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,10240,128,0.006236444330877728
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,256,2560,0.00656088896923595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,256,256,0.0022533333135975734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,2560,0.006950221955776215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,3584,10240,0.01590488851070404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,3072,2048,0.005602666487296422
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,2048,0.006575999988449945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,8192,0.014486221803559197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,10240,0.013788444300492605
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,2048,0.014399111270904541
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,3584,768,0.017941332525677152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,8192,16384,0.054795556598239474
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,10240,256,0.041060444381501936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,16384,2560,0.013090666797426013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,4096,128,0.005970666805903117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,128,1536,0.0038888889054457345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,768,10240,0.03538399934768677
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,128,128,0.010000000397364298
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1536,128,0.004877333425813251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,1536,0.007255111303594377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,1536,0.003510222252872255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,1536,256,0.005557333429654439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,1024,0.0058319999112023254
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,1024,0.005223111146026188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,16384,6144,0.023219555616378784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,4096,7168,0.013864888913101621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,5120,0.03103999959097968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,1536,0.012096000214417776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3584,1536,0.005889777921968036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,1536,0.006165333506133821
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,64,2048,0.006569777925809224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,256,0.014111111561457316
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,4096,512,0.006296000132958095
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,2048,0.004538666870858935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,256,0.005147555636035072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,6144,0.009716444545321995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,512,3584,0.01538755496342977
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,1536,2560,0.006963555183675554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,256,0.006463999963468975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,65536,8192,0.25277333789401585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,65536,16384,0.6652844217088487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,10240,0.031097776359981958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,128,0.002492444382773505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,128,256,0.014764444695578681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,128,0.02346755564212799
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3072,16384,0.022996443841192458
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,65536,2.0693732367621527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,8192,6144,0.018834667073355783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,4096,10240,0.03350222110748291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,768,0.010492444038391113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,8192,0.29693955845303005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,256,3584,0.013813333378897773
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,4096,5120,0.010978666444619497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,8192,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,3584,0.00830488900343577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,65536,5120,0.7120249006483289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,65536,768,0.015790222419632804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,512,0.014151111245155334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,256,16384,0.04129688938458761
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,7168,128,0.0069671107663048645
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,8192,0.027111111415757075
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,4096,2560,0.0069324444565508105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,256,768,0.014783999986118741
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,256,7168,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,65536,5120,0.056736886501312256
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,2560,0.00488711107108328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,512,8192,0.03018933203485277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,4096,16384,0.23928356170654297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,2048,64,0.002179555594921112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,1536,512,0.013933334085676404
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,1536,0.027112889620992873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,5120,0.008678221868144141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,1536,0.005963555640644497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,2560,768,0.018608000543382432
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,12288,0.027274666561020747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,4096,0.007653333246707916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,512,768,0.0058871110280354815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,1024,0.025790222816997107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,256,0.0025475554996066624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,768,1024,0.01277511070171992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,512,0.004903111192915174
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,65536,0.05885599719153511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,512,7168,0.00553955551650789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3584,10240,0.014751111467679342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,10240,3072,0.009034666750166152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,1024,2048,0.008366222182909647
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,128,0.005261333452330695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,6144,8192,0.024646222591400146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,256,0.01720888912677765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,768,0.006933333145247565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,768,6144,0.009360888765917884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,128,1024,0.0031893334041039148
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2048,128,0.01272266689274046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,5120,0.005222222043408289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,65536,10240,0.18888355626000297
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,5120,0.01959822244114346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,3072,1536,0.018423999349276226
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,10240,7168,0.0572942230436537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,6144,0.008980444735950893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,4096,8192,0.015056888262430826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,768,0.005594666633341048
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,5120,32,0.002518222149875429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,64,10240,0.03142044279310439
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,64,0.005424889011515512
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,6144,6144,0.22548977533976236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,5120,8192,0.030590222941504583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2048,64,0.005215999980767568
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,768,3584,0.0075911109646161394
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,1024,768,0.031108445591396753
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,65536,256,0.057462222046322294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,10240,0.018965333700180054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,12288,0.00489066665371259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,16384,10240,0.22316177686055502
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,512,7168,0.009733333355850643
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,512,65536,0.09646399815877278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,3584,768,0.01632355484697554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,768,3584,0.01751377847459581
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,7168,768,0.005548444473081165
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,768,0.022056000100241765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,768,8192,0.06695199675030179
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,2048,0.02072177827358246
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,128,0.006600889066855113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,768,8192,0.010856889188289642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,5120,0.06179910898208618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,5120,12288,0.025614221890767414
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,512,8192,0.01035377797153261
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,8192,2048,0.0075084442893664045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,768,0.005723555468850666
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,8192,128,0.014093332820468478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,128,10240,0.013029333617952136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2048,5120,0.016509332590632968
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,2560,256,0.004899555610285865
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,512,0.004883555488453971
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,65536,1024,0.06052088737487793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,768,512,0.017264889346228707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,512,2560,0.013768889009952545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,768,0.005612444546487596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,1536,0.006518222391605377
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,65536,5120,0.04974311259057787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,1536,0.022913777165942725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,1024,2560,0.007291555404663086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3584,16384,0.03271555569436815
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,5120,0.00908799966176351
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,12288,768,0.006968888971540663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,512,512,0.0026008888251251644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,7168,256,0.016269332832760282
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,16384,768,0.013529777526855469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,10240,0.05626933442221748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,2048,0.008570666942331526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,768,0.01616533266173469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2560,5120,0.011190222369299995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,128,0.0021119999388853707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,3584,0.0070053330726093715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,32,256,0.005180444568395615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,5120,0.024638222323523626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,16384,0.012005333271291522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,256,8192,0.020337777005301583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,2560,256,0.006298666612969504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1024,7168,0.010004444254769219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,10240,0.6740853521558973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,128,3072,0.008572444319725037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,512,10240,0.01311644415060679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,5120,0.008621333373917473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,768,6144,0.005718222094906702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,4096,0.07675821913613214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,1024,8192,0.011699555648697747
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,16384,768,0.005884444548024072
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,7168,0.010027555955780877
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,6144,0.011990221838156382
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,6144,10240,0.03523822294341193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,512,7168,0.018575110369258456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,16384,0.005584888988071018
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2048,32,0.006581333362393909
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,128,12288,0.02104355560408698
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,4096,0.022284444835450914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,1024,7168,0.009731555150614845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,2560,7168,0.176910228199429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1024,2048,0.005698666804366642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,8192,65536,0.08954755465189616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3584,768,0.005952000204059813
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,32,0.0063697778516345555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,12288,0.013082666529549493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,128,64,0.0018613334331247543
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,6144,256,0.004895110925038655
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,65536,0.08067555560006036
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,5120,5120,0.022358222140206233
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,512,0.002530666689078013
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,6144,512,0.015456888410780164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,10240,0.0432293348842197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,4096,4096,0.009285333255926767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,512,4096,0.0058648888435628675
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,1536,1024,0.0035075553589397003
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,8192,3584,0.05953333112928602
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,128,2560,0.006638222270541721
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,128,3072,0.012777778009573618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,512,16384,0.01889155473974016
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2560,10240,0.024319110645188227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,32,0.29269777403937447
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,256,0.005599110904667113
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,768,65536,0.05826489130655924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,6144,512,0.0031911111954185697
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,1024,0.011377777490350934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,12288,6144,0.10780355665418838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2048,10240,0.024283554818895128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,2048,0.005895111295912001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,3072,12288,0.1674426661597358
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,12288,0.028532445430755615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,512,10240,0.021156445145606995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,256,7168,0.005211555709441503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,8192,0.013797333670987023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,10240,1536,0.038077334562937416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,10240,16384,0.05347377724117703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2560,768,0.005718222094906702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1024,2560,0.004878222114510006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,16384,0.022245332598686218
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,16384,0.027120888233184814
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,1024,768,0.006640000061856375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,12288,256,0.002534222271707323
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,12288,0.012770666844315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,4096,1024,0.006566222343179915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,128,0.00245066670080026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,6144,0.027131555808915034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,3584,0.010874666273593903
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,10240,2560,0.012395555774370829
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,128,0.005151999907361137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,65536,512,0.0151235560576121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,7168,1024,0.010417777631017896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,2560,0.009743111001120673
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,256,0.0031946665710873077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,256,65536,0.35364532470703125
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,1536,65536,0.6380382113986546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,7168,10240,0.035213334692849055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,7168,16384,0.025644444757037695
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,1536,64,0.005227555417352253
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,7168,128,0.00868800034125646
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,128,1536,0.012120000190205045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,8192,16384,0.026531555586391028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,2560,0.008999110923873054
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,10240,16384,0.06226133637958103
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,65536,0.06233688857820299
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3584,10240,0.026777777406904433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,8192,65536,0.12763467099931505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,16384,128,0.005631111148330901
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,32,2560,0.004548444516128964
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,512,3584,0.007679111427730984
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,8192,0.23259022500779894
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,768,0.005592000153329637
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,768,64,0.006240889016124938
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,64,2560,0.006567111031876669
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,2048,65536,0.07658400138219197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,4096,128,0.019139554765489366
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,1536,0.01240355521440506
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,8192,0.027114666170544092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,1024,16384,0.024616888827747766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,6144,0.011729777687125735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,12288,12288,0.06304088566038343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,512,768,0.0028400001012616684
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,2560,7168,0.013384888569513956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,5120,0.03250577714708116
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,32,0.0052382221652401825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,2560,1536,0.014129777749379476
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,512,16384,0.011159111228254108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,3584,1536,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,512,2560,0.004879111217127906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,128,2048,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,7168,65536,0.1180293295118544
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,768,0.021586666504542034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,2560,0.0063591111037466265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,1024,0.013766222529941134
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,4096,7168,0.01889333294497596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,12288,8192,0.31892887751261395
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,12288,4096,0.016166221764352586
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,3072,0.007648888561460707
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,3072,0.03206400076548258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,2560,0.03221866819593642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,512,0.00518844442235099
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,2048,128,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,3072,7168,0.0120319997270902
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,16384,6144,0.04272355635960897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,3584,5120,0.025434666209750708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,3072,8192,0.01315822203954061
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,2560,6144,0.009007111191749573
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,1024,0.027108444107903376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,12288,1024,0.006963555183675554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,12288,0.05309244328074985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,65536,0.02472355630662706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,10240,2048,0.007204444044166141
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,128,0.006237333433495627
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,65536,32,0.004204444587230682
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,10240,5120,0.019704888264338177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,16384,0.032835556401146784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3584,512,0.008220444122950235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,64,0.0021955555097924336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,12288,7168,0.04870577653249105
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,10240,64,0.016939555605252583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,8192,6144,0.020308444897333782
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,12288,0.07604977819654676
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,4096,6144,0.03384266628159417
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,1536,2560,0.01816177699300978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,3072,0.020027556353145175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,2560,4096,0.02463200026088291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,6144,256,0.005577777822812398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,12288,256,0.004885333279768626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2560,16384,0.02260711126857334
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,64,6144,0.004888888862397936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,65536,256,0.012030222349696688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,2048,8192,0.009324444664849175
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,8192,2560,0.009619555539555019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,2048,32,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,12288,12288,0.12522399425506592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,3072,16384,0.2024071084128486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,10240,0.03070488903257582
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,8192,16384,0.05907289187113444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,64,1024,0.006303999986913469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,8192,1024,0.023183110687467787
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,256,6144,0.029783109823862713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,8192,768,0.004273777620659934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,3584,0.019326221611764696
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2560,2048,0.03187911046875848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,1536,0.005895111295912001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,10240,1024,0.010431999961535135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,768,1536,0.006312888943486744
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,6144,4096,0.0244168887535731
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,12288,1024,0.008672888908121321
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,64,12288,0.014112000664075216
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,4096,1024,0.015492444237073263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2560,256,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,2560,0.01976355579164293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,512,1024,0.012720000412729053
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,256,6144,0.017892445127169292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,64,0.002100444502300686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,2560,1024,0.003583999971548716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,256,32,0.006568000134494569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,128,3072,0.007322666545708974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,512,0.002536888958679305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,128,256,0.005786666853560342
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,128,0.014473777678277759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,7168,32,0.005863111052248213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,8192,0.013765333427323235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,16384,32,0.0028328889360030494
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,5120,65536,0.08481066756778294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,2560,768,0.01717066599263085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,7168,0.023369777533743117
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,4096,2048,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,65536,0.060292442639668785
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,12288,4096,0.05063822203212314
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,16384,1536,0.02810133496920268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,5120,0.03085511260562473
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,1024,0.006272000157170826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2560,6144,0.01092266705301073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,16384,2560,0.016542222764756944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,3072,256,0.004931555440028508
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,7168,0.019101333287027147
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,1024,10240,0.11918132834964329
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,5120,1024,0.005590222362014983
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,12288,12288,0.035374220874574445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,3072,16384,0.019842666056421068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,512,128,0.0021475555582178966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,5120,10240,0.01651822196112739
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,65536,1024,0.04512088828616672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,768,0.0028213332924577924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,3072,10240,0.03806133402718438
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,512,2560,0.004924444274769889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,4096,5120,0.1878782245847914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,6144,5120,0.014479110638300577
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,2048,32,0.0052577778697013855
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,3584,12288,0.01938222183121575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,256,0.011094222466150919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,16384,16384,0.03411111235618591
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,3072,1024,0.003192000091075897
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1536,512,0.0026586666289303037
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,64,128,0.0021564445147911706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3584,65536,0.10269688897662693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,2560,1536,0.005965333431959152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,6144,6144,0.02624622152911292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,8192,32,0.0018479999982648427
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,2048,0.00627377794848548
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,3072,3072,0.007276444799370236
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,10240,2560,0.09100711345672607
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,2560,0.007241778075695038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,3584,6144,0.01039644413524204
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,4096,3072,0.007264888948864407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,16384,10240,0.05447999967469109
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,2560,65536,0.13866578208075628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,16384,0.02953688965903388
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,512,256,0.0024986666523747975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,32,768,0.005647111270162795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,65536,256,0.007931555310885111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,512,12288,0.01309511148267322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,16384,6144,0.04042399923006693
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,512,0.011732444167137146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,5120,0.023128888673252527
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,512,1536,0.0035573335157500375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,768,0.014545778433481852
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,768,10240,0.012297777665985955
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,8192,6144,0.01457333399189843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,16384,2048,0.012739555703269111
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,8192,10240,0.053948445452584155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,3072,4096,0.1200408935546875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,32,12288,0.004941333499219683
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,3072,0.011398221883508893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,10240,2048,0.010103999740547603
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,3072,32,0.0018542222678661346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,8192,3584,0.011353777514563667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,1536,1536,0.014103111293580798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,4096,0.012120000190205045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,5120,12288,0.03532800078392029
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,3072,0.025776000486479864
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,128,32,0.00554577757914861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,12288,6144,0.06809955835342407
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,12288,3072,0.016287111573749118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,32,0.005239999956554837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,5120,3072,0.022294221652878657
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,2048,0.013437333206335703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,16384,16384,0.7471884621514214
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,768,512,0.003112888791494899
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,1024,256,0.0141004439857271
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,65536,7168,0.049590223365359835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1536,256,0.00654755574133661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,256,3072,0.015440000428093804
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,12288,8192,0.017980444762441847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,65536,1024,0.012107555237081317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,3584,64,0.00591111100382275
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,128,2048,0.0127688886390792
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1536,768,0.016922665966881644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,32,0.005244444641802046
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,16384,65536,0.21721333927578398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,3584,0.005920889063013925
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,8192,256,0.007928888830873702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,512,1024,0.013024888932704926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,10240,5120,0.026727999250094097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,12288,2048,0.010683555569913654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,7168,3584,0.11328621705373128
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,1536,65536,0.45588000615437824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,6144,3072,0.03141333328353034
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,10240,1024,0.008608889248636033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,1024,0.005632888939645555
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,12288,1024,0.03375466664632162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,768,128,0.010372444159454769
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,7168,6144,0.014562666416168213
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,5120,12288,0.04075377848413255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,1024,0.015837333268589444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,5120,0.01166133334239324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,32,0.005552000055710475
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,3072,2048,0.006599111275540457
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,512,3072,0.007277333074145847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,512,0.015105777316623263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1024,8192,0.0255084451701906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,3072,0.0069520001610120135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,8192,0.014870221416155497
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,1536,4096,0.008623999853928884
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,10240,0.011004444625642566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,2048,0.018578666779730055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,4096,128,0.0028560000161329904
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,128,7168,0.004919111314747068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,2560,8192,0.01791288952032725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,10240,0.012219555675983429
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,4096,512,0.005577777822812398
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,7168,0.24891021516588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,12288,512,0.006659555352396435
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,12288,12288,0.2969706588321262
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,8192,0.11078400082058376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,2048,0.026473777161704168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,128,0.005895111295912001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2560,7168,0.00982844415638182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,768,7168,0.022482666704389784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,768,4096,0.01536355581548479
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,128,2048,0.012072000238630505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,10240,512,0.03650577863057455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,256,256,0.004587555511130227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,256,16384,0.014831999937693277
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,10240,3584,0.03518666823705038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,2560,0.004908444566859139
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,32,64,0.0018577778504954444
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,65536,1024,0.027619555592536926
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,65536,0.007308444215191736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,8192,1536,0.00700711127784517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,10240,6144,0.040468444426854454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,8192,256,0.045478221442964345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,4096,8192,0.02479466630352868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,2560,1536,0.016143111719025504
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,10240,2048,0.0075839997993575195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,64,8192,0.005264889034960005
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,8192,0.018163555198245578
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,16384,0.2578168974982368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2560,16384,0.11296978261735703
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,3072,768,0.00544888898730278
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,16384,1024,0.008345777789751688
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,7168,0.04219733344184028
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,64,0.0018097776919603348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,768,0.002480000050531493
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,2048,1536,0.010882666541470421
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,6144,0.008296888735559251
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,7168,8192,0.029948443174362183
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,3584,3584,0.008953777452309927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,512,0.005623999983072281
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3584,3584,0.008620444271299574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,65536,1536,0.0675519969728258
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1536,4096,0.02021422154373593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,2560,12288,0.028140445550282795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,16384,64,0.002526222210791376
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,768,10240,0.011784888803958893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,6144,16384,0.10143199894163345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,1024,12288,0.010415111151006486
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,10240,3072,0.04082399937841628
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,2048,12288,0.019394665956497192
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,65536,0.12984266546037462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,512,65536,0.07284533315234713
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,65536,0.047416889005237155
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,8192,0.017841777867741056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,12288,2048,0.0086986662613021
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,2048,5120,0.009648888475365108
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,1024,0.00556355549229516
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,4096,128,0.013799111048380533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3072,4096,0.013167999684810638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,12288,2560,0.010053333308961656
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,3584,2048,0.006606222026877933
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1536,2560,0.005094222310516569
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,2048,0.0281333327293396
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,32,128,0.002136888810329967
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,1536,0.021606221795082092
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,3584,7168,0.019633778267436557
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,1536,32,0.0021520000365045336
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,64,128,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,7168,6144,0.01238933371173011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,10240,0.01900800069173177
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,65536,2048,0.07717955774731107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,8192,12288,0.03173155585924784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,64,2048,0.006817777951558431
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,4096,6144,0.012723555167516073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,5120,0.02780533168050978
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,8192,2560,0.08481866783565944
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,10240,0.13523022333780924
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,32,768,0.005906666732496685
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,4096,0.008533333738644918
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,7168,65536,0.0847288899951511
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,1024,2048,0.015173332558737861
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,64,1536,0.003773333297835456
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,32,16384,0.0052533331844541765
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,1536,12288,0.01592266725169288
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,32,5120,0.008882666627566019
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3584,7168,0.012693333129088083
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,65536,2048,0.02843377656406826
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,65536,5120,0.036534223291609026
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,16384,2.500104056464301
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,65536,768,0.04628800021277534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,10240,10240,0.039628442790773176
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,8192,16384,0.04222044348716736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,1024,6144,0.009101333717505137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,2048,3584,0.016174222032229107
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,2560,0.017920000685585868
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,1024,2048,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,2048,3072,0.007311999797821045
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3584,3072,0.017321777012613084
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,64,12288,0.005607999861240387
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,5120,1536,0.00629333323902554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,768,12288,0.011928000383906893
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,12288,7168,0.03172977765401205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,16384,64,0.0021342221233579847
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,65536,64,0.010011555420027839
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,8192,65536,0.10089511341518825
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2048,8192,0.00942488925324546
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,8192,0.01407999959256914
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,768,16384,0.0188071115149392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,4096,32,0.0021946666141351065
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,10240,2048,0.027473777532577515
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,65536,5120,1.8520942264133031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,2048,1024,0.018256889449225534
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,12288,2048,0.01237155579858356
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,80,6144,2560,0.006934222247865465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,8192,3584,0.011393778026103973
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,8192,0.020949333906173706
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,65536,128,0.015846222639083862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,8192,6144,0.03317777646912469
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,12288,7168,0.03590666585498386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,2048,0.00629688882165485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,5120,7168,0.026644445127911035
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,1536,4096,0.007251555720965068
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,1024,2048,0.01682133310370975
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,6144,16384,0.03874666823281182
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,256,8192,0.018797333041826885
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,3584,1536,0.005936000082227919
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,256,5120,0.009003555609120263
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,5120,0.005932444499598608
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,4096,0.004948444250557157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,16384,3584,0.03705688979890611
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,3072,3072,0.007112888826264276
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,256,7168,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,512,4096,0.04479733440611097
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,4096,1024,0.003552000141806073
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,16384,4096,0.35175734096103245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,768,0.0031715554909573663
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,1024,768,0.007329777710967594
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,2048,0.01514488955338796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,256,65536,0.01626311077011956
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,256,3072,0.007268444531493717
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,10240,8192,0.024069334069887798
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,512,65536,0.06621866756015353
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,2560,5120,0.03512800070974562
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,256,3072,0.005260444349712796
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,6144,6144,0.01901333365175459
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,1024,65536,0.05919466416041056
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,12288,12288,0.04633333285649618
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1024,3072,0.009700444837411245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,12288,1536,0.07202488846249051
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2560,6144,0.01146755533085929
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,5120,3584,0.00868088917599784
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,5120,1024,0.016164445214801364
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,4096,1024,0.014881778094503613
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,7168,0.013718222578366598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,4096,64,0.0021173333128293357
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,6144,3584,0.009630222287442949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,128,7168,0.018584888842370775
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,64,256,0.002522666628162066
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,1536,0.009726222190592024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,512,128,0.0025191110455327565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,3072,16384,0.03386933273739285
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,3584,0.009752000371615091
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,256,12288,0.012734221915404001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,6144,3584,0.0086666668454806
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,12288,7168,0.03353510962592231
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,10240,3072,0.013247999880048962
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,2560,768,0.011528888510333167
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,1536,4096,0.007890666524569193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2,7168,2048,0.006592888798978593
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3072,3584,0.009677333136399588
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,128,2048,0.004587555511130227
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,6144,0.014519110321998596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,256,6144,0.009312888814343346
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,256,128,0.012446222205956778
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,2048,10240,0.01478666729397244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,6144,32,0.0028657778683635923
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,256,512,0.011030221978823343
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,65536,5120,0.7961280081007215
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1536,10240,0.016525333126386006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,1536,32,0.002157333410448498
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,7168,16384,0.027186666925748188
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,64,0.001840888833006223
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,3584,768,0.0058755555914507965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,256,0.017242666747834947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,8192,3584,0.011983999775515662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,12288,0.01962844365172916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,1536,10240,0.010335110955768162
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,2048,1024,0.0162426663769616
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,768,0.0028204443968004654
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2048,2560,0.007119110888904995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,1024,6144,0.03194933467441135
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,8192,3584,0.030304001437293157
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,2048,12288,0.016591111818949383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,16384,5120,0.05367822117275662
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,12288,3072,0.03054666519165039
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,3584,2560,0.04175733195410835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,4096,2048,0.022964444425370958
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1024,16384,1536,0.021182222498787776
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,128,4096,0.0038497779104444715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,64,0.027983112467659846
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,1536,0.019241778386963736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,3072,4096,0.02608088817861345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,2048,4096,0.008347555167145198
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,1024,2048,0.006278222219811545
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,8192,0.02885155545340644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,12288,65536,0.16188355286916098
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,12288,64,0.03426933288574219
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4096,512,10240,0.03175110949410333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,7168,0.05429244372579786
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,1536,7168,0.009659555223253038
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,768,0.012741333080662621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,6144,1024,0.016552888684802584
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,16384,1536,0.061786664856804736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,7168,768,0.003569777641031477
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,65536,7168,0.1542159981197781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,16384,0.017886221408843994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,6144,0.017567111386193168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,6144,512,0.0069777775141927935
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,2048,2560,0.007170666423108842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,3072,256,0.002568888788421949
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,10240,10240,0.0399635566605462
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,5120,5120,0.013089777694808112
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,5120,1024,0.009359111388524374
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,1024,2048,0.016531555189026725
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,512,1024,0.012031111452314587
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,12288,10240,0.06594400273429023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,65536,128,0.007304888632562425
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,6144,0.027152000202072993
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,6144,3584,0.15145244863298205
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,5120,256,0.0028880000528362063
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,64,65536,0.006629333313968446
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,384,3584,7168,0.01866311166021559
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,1536,3072,0.01516533394654592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,12288,768,0.005927111125654644
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,10240,0.23151199022928873
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,768,1536,0.016167110866970487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,7168,64,0.0021404443929592767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,3072,16384,0.0472924444410536
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,256,7168,0.009680888719028896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,1024,3584,0.005200888961553574
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,1024,1536,0.0038417776425679526
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,65536,16384,1.2538746727837455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,16384,64,0.011340444286664328
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,768,3584,0.009647111097971598
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,5120,1536,0.07876355780495538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,64,64,0.0022088889446523455
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,64,12288,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,1024,2048,0.01848355597919888
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,7168,512,0.017218665944205392
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,5120,7168,0.025556445121765137
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,8192,7168,0.017539555827776592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,2048,0.006310222049554189
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,768,2560,0.0069511110583941145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,4096,0.0076142218377855085
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,5120,5120,0.016677333248986136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,1536,1536,0.003867555409669876
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,64,4096,0.007628444168302748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,10240,3584,0.013609777722093793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,5120,8192,0.05590666664971245
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,8192,128,0.017673777209387887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,16384,1024,0.007074666519959767
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,128,7168,0.024711999628278945
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,3072,512,0.002489777695801523
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,2048,1024,0.006243555496136348
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,1024,0.005240889059172736
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,512,10240,0.03151466780238681
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,2048,7168,0.013608000344700284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,7168,1536,0.017863111363516915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,512,16384,0.008343111309740279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,512,4096,0.03120533294147915
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,65536,0.13425421714782715
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,2560,32,0.001963555605875121
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,3072,2048,0.006634666687912411
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,6144,2048,0.019878221882714164
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,6144,256,0.014507555299335055
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,512,64,0.0021377777059872947
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,3584,32,0.004847110973464118
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,5120,64,0.005224888937340842
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,16384,12288,0.27356799443562824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,128,5120,0.009017777939637503
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,2048,0.004471999903519948
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,10240,12288,0.045046223534478076
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,128,1024,0.005944889038801193
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,2560,0.007991111112965478
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,1536,1024,0.0059279998143513995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2048,768,0.005283555636803309
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,256,6144,0.01721244388156467
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,32,1024,0.002871111035346985
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,16384,512,0.09165777762730916
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,3072,3072,0.021893334057595994
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,1024,128,0.00249422217408816
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,16384,0.1094506647851732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,2560,0.007306666837798224
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,10240,0.0244195560614268
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,1024,5120,0.008656000097592672
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,256,256,0.0052560000783867305
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,2048,32,0.0018275556051068837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,512,7168,0.011402666568756104
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,1536,1536,0.0063075555695427795
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,3584,0.020611556039916146
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,4096,2560,0.02219733264711168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,2048,12288,0.11946845054626465
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,65536,7168,0.2552808920542399
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,10240,16384,0.060589333375295006
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,16384,512,0.030829333596759375
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,8192,1536,0.006287999865081575
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,4,64,16384,0.00553955551650789
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,8192,0.014799111419253878
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,3072,5120,0.010232888989978367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,7168,1536,0.020624000165197585
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,16384,3584,0.058279110325707324
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,256,0.006284444282452266
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,5120,2560,0.022639110684394836
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,768,256,0.0025048889219760895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,256,65536,0.09464889102511936
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,2560,0.011197333534558615
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,128,65536,0.07611911164389716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,3584,256,0.005387555393907759
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,3584,1024,0.024566221568319533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,65536,1536,0.10393066538704766
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,8192,4096,0.057031108273400195
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,12288,2048,0.011206222077210745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,512,512,0.0026133333643277488
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,10240,1024,0.0059164443777667145
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,3072,7168,0.011026666396194033
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,1536,10240,0.012431999875439538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,128,0.0021422221842739317
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,256,6144,0.009534222384293875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,256,512,65536,0.06231111288070679
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,16,1536,3072,0.0052471111218134565
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,3072,0.02420799930890401
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,64,8192,0.011008888483047485
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,5120,7168,0.036673777633243136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,2048,2048,0.03788622220357259
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,5120,65536,0.08105333646138509
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,768,2560,0.014532445205582513
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,256,1024,0.012048888537618848
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,4096,10240,0.017501332693629794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,768,65536,0.026309332913822595
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,3072,1024,0.006153777655627992
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,7168,768,0.03221866819593642
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,16384,0.03551910983191596
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,768,7168,0.01953866746690538
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,4096,7168,0.05333422289954292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,128,8192,0.04387377699216207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,4096,4096,0.012116444607575735
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,65536,3584,0.09990044434865315
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,4096,2560,0.017516444126764934
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8192,256,7168,0.058768888314565025
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,256,65536,0.07705244753095838
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,2048,3072,0.03345688846376207
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,5120,2560,0.007291555404663086
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,8192,12288,0.037351999017927386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,64,5120,0.006597333484225803
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,10240,12288,0.03502577874395583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,3584,128,0.013756444056828817
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,7168,64,0.0021297778520319196
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,6144,128,0.013799111048380533
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,80,768,4096,0.015795555379655626
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,8192,512,0.0069297779765393995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,192,1536,2560,0.015182221929232279
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,6144,1536,0.006256000035338932
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8192,7168,128,0.034239110019471906
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,256,10240,0.011745777395036487
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,768,6144,0.009695111049546136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,7168,8192,0.03132177723778619
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,10240,16384,0.11422667238447402
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,512,1024,0.006960888703664144
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2,128,10240,0.005251555393139522
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,3072,8192,0.03341066506173875
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,8192,2560,0.028384000062942505
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,160,256,5120,0.006576889091067844
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,128,128,0.004581333448489507
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,512,8192,0.020609777834680345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,2560,64,0.005581333405441708
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,256,4096,0.014145778285132514
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2,3584,4096,0.027767111857732136
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,2560,1536,0.007350222104125553
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,65536,6144,0.12461422549353705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,512,65536,0.9846622678968641
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,7168,2560,0.007215110792054071
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,2048,5120,0.011191111471917896
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,65536,3072,1.8503733740912542
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,6144,5120,0.02409511142306858
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,128,512,0.006549333532651265
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,12288,32,0.007273777491516537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,768,128,0.0069226668112807805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,8192,3072,0.011024889018800525
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,5120,8192,0.05472266674041748
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,256,6144,4096,0.03267200125588311
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,768,128,0.004577777865860197
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,768,512,0.0024826667375034755
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,3072,12288,0.020990222692489624
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,12288,1024,0.09327822261386448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,3584,2048,0.006507555643717448
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,512,8192,0.02331111166212294
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,65536,3584,0.04209422071774801
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,2560,4096,0.014646222194035849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,64,128,1024,0.0028577778074476454
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,128,2048,0.006231110956933763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,2560,1024,0.005560889012283749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,1536,256,0.005720888988839255
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,6144,65536,0.12833155526055232
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,6144,256,0.016220443778567843
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2,10240,3584,0.013168888787428537
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,768,7168,0.01792355544037289
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,2048,16384,0.05014666583802965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,8,7168,256,0.0052782222628593445
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,2560,12288,0.03808266586727566
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,32,2048,1536,0.0036391110883818734
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,3584,256,0.0037235555549462638
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,128,2048,0.012661333713266583
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,1024,3072,0.008011555506123437
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,1,8192,6144,0.012741333080662621
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1,256,3584,0.013805333111021252
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,512,1536,5120,0.009361777868535783
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,32,5120,0.009301332963837517
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,512,2560,0.006769777586062749
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,16,5120,8192,0.01484266585773892
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,2560,65536,0.05431822273466322
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,768,7168,0.009699555734793345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,128,12288,16384,0.045062220758861966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,512,1024,4096,0.01162577751610014
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,256,65536,0.07339733176761203
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,512,256,16384,0.0247013337082333
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,768,128,0.004909333255555895
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,16,8192,10240,0.03427555494838291
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1,3584,65536,0.07368355327182345
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,1024,8192,0.01237777786122428
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,80,256,512,0.005593777944644292
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2048,256,0.005971555494599872
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,1024,7168,0.020495999190542433
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,64,768,0.0026346666531430352
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,16,4096,1536,0.006270222365856171
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,128,3584,0.004936000125275718
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,6144,6144,0.07202311356862386
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,5120,1536,0.007632888853549957
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,2048,7168,256,0.007987555530336168
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,64,7168,1024,0.017384888397322763
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,32,256,0.0020933333370420667
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,16384,10240,0.0237128883600235
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,256,0.004943111290534337
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,128,1024,0.005606222069925732
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,1536,128,0.014139556222491793
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,2048,5120,3072,0.044897778166664966
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8192,3584,3584,0.10847022136052449
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,2048,128,3072,0.0107022225856781
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,128,128,768,0.012086222569147745
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,2560,1536,0.00658044425977601
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,48,2560,16384,0.02331466641690996
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,384,7168,7168,0.03698933455679152
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,32,10240,6144,0.03085955646302965
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,128,128,0.0018622221218215094
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4,2560,3584,0.007499555746714274
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,5120,768,0.0039048890272776284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,1536,7168,0.013375110924243927
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,2560,4096,0.00795377790927887
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,3072,768,0.016549333930015564
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,7168,4096,0.039157334301206805
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,96,256,12288,0.006561777657932705
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,6144,1024,0.005904888941182031
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,256,2560,0.014372444815105863
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,2048,256,65536,0.17531466484069824
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,256,256,2560,0.004893333133724001
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,48,512,6144,0.01682844426896837
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,768,12288,768,0.011239999698268043
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,8192,2048,0.031938665442996554
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,32,3072,12288,0.01885333326127794
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,12288,6144,0.05062222149637011
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4,512,4096,0.01681688924630483
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,80,65536,8192,0.17032710711161295
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,768,2048,0.014730667074521383
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,48,16384,512,0.00452533322903845
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,48,1024,12288,0.013757333159446716
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,64,2560,0.004867555366622077
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,32,2560,7168,0.01236533373594284
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,512,1024,0.005596444424655702
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,768,10240,768,0.012785777449607849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1024,512,65536,0.12675466802385119
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,192,512,4096,0.00830577810605367
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,384,5120,4096,0.014741332994567024
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,96,5120,1536,0.006585777633719974
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,64,768,768,0.00535022219022115
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,64,128,768,0.005616888817813661
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,8,6144,7168,0.04028622309366862
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8192,7168,5120,0.3237440056271023
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,384,3072,512,0.005559999909665849
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,96,5120,128,0.00518133325709237
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,160,65536,2048,0.034076445632510714
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,8,10240,1024,0.006098666538794835
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,128,65536,2560,0.02369955513212416
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4,4096,65536,0.07874933216306898
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,1,1024,65536,0.055381334490246244
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,160,3072,2048,0.006610666712125142
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,192,64,6144,0.004914666629499859
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,1024,512,512,0.012753778033786349
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,4096,8192,3584,0.09823377927144368
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,256,65536,8192,0.07785422272152372
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,192,12288,128,0.006242666807439592
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,1024,2560,7168,0.02460533380508423
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8,4096,6144,16384,0.3126640054914686
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,768,8192,65536,0.4816382196214464
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,160,12288,128,0.0173973325226042
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,float16,128,65536,1536,0.0403164459599389
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,96,7168,5120,0.02862666712866889
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,8,7168,1536,0.005943999936183293
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,16384,768,0.04035733474625482
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,nvfp4,4096,65536,3072,0.33388355043199325
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,512,1024,5120,0.020483555065260995
SGLang,0.5.9,NVIDIA GB300,gemm,sglang,fp8_block,768,2560,7168,0.03468177715937296
